The CFO's Guide to Tariff Exposure Modeling: From Cash-Flow Stress Tests to Board Reporting

By the end of 2025, CBP had collected roughly $133.5 billion in duties under a single tariff authority, the International Emergency Economic Powers Act (IEEPA). In February 2026, the Supreme Court ruled 6-3 that those tariffs were imposed unlawfully. The money is now being refunded through CBP’s CAPE system, slowly, entry by entry, while the government appeals parts of the refund order.

For a finance chief, the problem is that cash goes out the door on a fixed schedule, and it comes back (if it comes back) on a timeline nobody controls. This has led to a working-capital gap that most companies never modeled, on the assumption that duties were small enough to bury inside landed costs. In this blog, we’ll walk through why that assumption no longer holds up at current rates. 

Why Tariff Exposure is Now a CFO Issue

The shift from tariff line-item to enterprise risk

For years, tariffs were a customs problem. A broker classified the goods, a duty rate applied, the number landed in cost of goods sold, and finance rarely looked closer. But now that duty rates on a category can move from 25% to 50% in a single proclamation, it’s introduced volatility that pulls tariffs onto the CFO’s desk. A cost that swings with litigation and executive action affects finance responsibilities like margin forecasts and cash planning. 

Materiality thresholds in 2026's multi-layer duty stack

One reason the numbers have grown is that duties now stack. A single imported product can carry a most-favored-nation base rate (MFN, the standard duty owed by normal trading partners), plus a Section 232 metals tariff, plus a Section 301 rate tied to Chinese-origin content, plus antidumping and countervailing duties (AD/CVD) if the goods fall under an active order. Each layer is set under a different statute, on a different timeline, by a different agency.

Stacked together, these layers can push the effective duty on a product well past the point where it moves a segment's margin. Exposure that once sat below any reasonable materiality threshold can cross it without a single change to your product line, purely because a new layer attached to goods you were already importing.

Board oversight expectations under recent SEC guidance

Regulators have noticed the same shift. Guidance heading into the 2025 and 2026 reporting seasons has pushed companies to quantify tariff effects where material, describe mitigation efforts, and treat evolving trade policy as a known trend or uncertainty rather than background noise. Even companies not yet materially exposed are expected to monitor and, where appropriate, disclose the risk of future volatility. In short, CFOs need defensible answers.

Building the Tariff Exposure Model

Per-SKU duty stack (MFN + 122 + 301 + 232 + AD/CVD)

A credible exposure model starts at the SKU level, because that is where the stack actually resolves. For each product, the model should carry every applicable layer: the MFN base, any Section 122 surcharge, Section 301 rates on Chinese content, Section 232 metals duties, and any AD/CVD orders the goods fall under. The layers are additive, and they apply against different values, so a single blended rate hides more than it reveals.

Sourcing-country sensitivity analysis

Once the stack is built per SKU, the next question is how sensitive each line is to where it comes from. Country of origin drives most of the layers: Section 301 keys off Chinese origin, AD/CVD orders name specific countries, and Section 232 now turns partly on where metal was melted and poured. A model that can flex origin shows you which products get cheaper or more expensive under a sourcing change before you commit to one, which matters because a sourcing decision now carries a duty consequence that can dwarf unit cost and lead time.

Scenario modeling for active regulatory events (Section 122, Solar IV, and IEEPA)

The layers are not static, so the model should not be either. Several duty sources are in open motion right now. The 10% Section 122 global surcharge was struck down by the Court of International Trade in May 2026, but the ruling was stayed on appeal and the surcharge itself is set to expire in late July 2026, so its status could change more than once in a quarter. The AD/CVD case on solar cells from Cambodia, Malaysia, Thailand, and Vietnam (widely called Solar IV) continues to expand, with newer investigations reaching additional countries. And the IEEPA refunds are still working through the courts.

Scenario modeling means running the exposure number under each plausible outcome: surcharge expires or is reinstated, an order widens, a refund lands or slips. The output is not a forecast so much as a set of bounded cases finance can plan against and, when asked, defend.

Cash-Flow Stress Tests

Working capital impact of duty deposits vs. refund timing

The working-capital point is the one most models miss. Duties are paid as deposits at entry, on a predictable schedule tied to your import volume. Refunds, whether from a favorable ruling or a corrected filing, arrive on a schedule set by agencies and courts. The gap between the two is real cash, tied up for months or longer, and it belongs in a stress test the same way a receivables slowdown would.

Liquidation timing assumptions

Part of that gap comes from liquidation, the point at which CBP finalizes the duties owed on an entry. Until an entry liquidates, the amount is not settled, and refund or additional-payment outcomes are not fixed. Realistic stress tests carry explicit liquidation-timing assumptions rather than treating deposits as final at entry, because the difference between "paid" and "settled" can span a year or more.

Protective protest receivables as contingent assets

When a company believes it overpaid, it can file a protest to preserve its right to a refund while the underlying question is resolved. Those protective protests create a potential receivable, but a contingent one: the cash depends on an outcome that has not happened yet. Treating protest-based recoveries as contingent assets, rather than as money already earned, keeps the stress test honest and keeps the board from planning around cash that may not arrive.

Four Reportable Metrics for the Board

Boards do not need the full model. They need a small set of numbers that move together and tell a coherent story. Four cover most of the ground.

Gross tariff expense (current period)

The first is gross tariff expense for the period: total duties actually incurred, before any recovery. This is the headline exposure number, the one that shows how much the current trade environment is costing the business right now.

Refund receivable (IEEPA CAPE + protective protests)

The second is the refund receivable, split between claims moving through CBP's IEEPA refund system (the module CBP built inside ACE, called CAPE, the Consolidated Administration and Processing of Entries) and recoveries tied to protective protests. Because refunds are not automatic and are still partly contested in court, this figure should be reported with its collection confidence, not as a certainty.

Classification reserve (audit and dispute contingency)

The third is a classification reserve: an amount set aside against the possibility that some entries were classified incorrectly and additional duty is owed. This is the counterweight to the refund receivable, and it deserves equal attention. In a Gaia study of 30,000 imported line items, roughly a third carried discrepancies, which is a useful reminder that recovery and exposure often live in the same set of entries.

Supplier-renegotiation pipeline (forward mitigation)

The fourth looks forward: the pipeline of supplier and sourcing renegotiations underway to reduce future duty. Reported alongside the backward-looking numbers, it shows the board how much of the current expense is structural versus addressable.

Tariff Disclosure in 10-K and 10-Q

When tariffs cross the materiality threshold

Disclosure obligations turn on materiality, and the stacking problem is what makes materiality a moving target. As new layers attach, tariff cost can cross from immaterial to material between filings, so the assessment needs to be redone each period rather than settled once.

MD&A discussion of tariff impact

In Management's Discussion and Analysis (MD&A), recent guidance favors quantified detail: the rates applied, the effect on margins, the sourcing mix behind the exposure, and the mitigation steps in motion. Qualitative gestures toward "tariff headwinds" no longer meet the expectation when a real number is available.

Risk-factor language for active litigation

Because several tariff sources are in active litigation, risk-factor language has to hold two things at once: the possibility that duties are reduced or refunded, and the possibility that they are reinstated or expanded on appeal. Drafting for the uncertainty itself, rather than betting on one outcome, is the safer posture while cases like the IEEPA and Section 122 challenges remain open.

Contingency accounting for protective protests

The protest receivables that appear in the stress test also carry an accounting treatment. Because the recovery is contingent on an unresolved outcome, it is generally accounted for as a contingency rather than recognized as a firm asset, and the disclosure should make the conditional nature of the recovery clear.

Pricing Pass-Through Decisions

Modeling customer-side absorption capacity

Every duty dollar is absorbed by the company, passed to customers, or split. Which happens depends on how much price your customers can take, and that varies by segment, contract structure, and competitive position. Modeling absorption capacity per customer or channel turns pass-through from a blanket decision into a targeted one.

Timing pricing adjustments to tariff events

Timing matters as much as amount. Because duty rates move on policy events, a price increase taken just before a surcharge expires can misfire, and one taken too late leaves margin on the table. Tying pricing moves to the same scenario calendar that drives the exposure model keeps the two decisions in sync.

Margin impact of partial vs. full pass-through

The choice between partial and full pass-through is a margin question with a demand consequence: full pass-through protects margin but risks volume, while partial pass-through defends volume at the cost of margin. Running both through the model, by segment, gives finance the tradeoff in numbers rather than instinct.

Building a Tariff Exposure Dashboard

Data inputs from ERP, broker, and CBP entry data

A dashboard is only as good as its inputs, and tariff exposure draws from three systems that rarely talk to each other: the ERP (product, cost, and sourcing data), the broker's classification and entry records, and CBP's own entry and liquidation data. Bringing them into one view is most of the work, because discrepancies between them are often where hidden exposure lives.

Update cadence (real-time vs. monthly)

Not every input needs to refresh at the same speed. Entry-level activity and duty deposits benefit from a near-real-time feed, while scenario assumptions and board metrics are fine monthly. Matching the refresh rate to how fast each number actually moves keeps the dashboard trustworthy without over-engineering it.

Shared ownership across Finance, Compliance, and Sourcing

The dashboard also needs an owner, and tariff exposure does not sit cleanly with any one team. Finance owns the cash and disclosure implications, compliance owns classification accuracy, and sourcing owns the origin decisions that drive the stack. Shared ownership, with finance accountable for the reported numbers, tends to work better than assigning the whole thing to one function that only sees part of the picture.

How Gaia Dynamics Fits

Most of the difficulty here is data: getting an accurate per-SKU duty stack, keeping it current as layers change, and checking historical entries before you act on them. Gaia's Tariff Engine models the full duty stack per product and runs the kind of scenario cases described above as regulatory events move. Its Tariff Audit feature checks entries for the classification discrepancies that create both refund opportunities and exposure, which is the review that belongs before any refund claim, not after.

If your team is building a tariff exposure model or preparing a refund position, you can put your own product and entry data through Gaia and see the stack, the scenarios, and the discrepancies for yourself. [Get started for free.]

FAQs

When does tariff exposure become material for SEC reporting? There is no fixed dollar threshold; materiality depends on the effect relative to your results and the total mix of information available to investors. Because duties now stack across multiple statutes, exposure can cross into material between filings without any change to your products, so the assessment should be redone each reporting period.

How do I account for an IEEPA refund receivable? Refunds under CBP's IEEPA process are not automatic and remain partly contested on appeal, so recovery is generally conditional rather than assured. Many companies treat these as contingent until the claim is validated and the collection outcome is reasonably certain, and disclose the conditional nature rather than booking a firm asset early. This is an accounting judgment worth confirming with your auditors.

Should tariffs sit in COGS or a separate line? Duties are typically capitalized into inventory and flow through cost of goods sold as those goods are sold, which is why tariff cost often hides inside blended landed cost. For internal reporting and board visibility, breaking tariff expense out as its own tracked figure makes the exposure legible even when the financial statements fold it into COGS.

How often should the board review tariff exposure? Given how fast duty sources are moving, quarterly review fits most companies, aligned to the reporting cycle, with an out-of-cycle briefing when a major event lands (a court ruling, a new order, or a surcharge expiring). The four board metrics above are designed to make that review quick.

What's the right accuracy benchmark for the model? The practical benchmark is the accuracy of a strong licensed customs broker, since that is the standard CBP holds importers to. Classification accuracy is the foundation: if the underlying HTS assignments are wrong, every downstream number, from exposure to refund to reserve, inherits the error.