2027 planning assumptions need two tariff cases, not one
The CFO Survey shows tariff-exposed firms internally modeling growth close to 1.6% while the broad sample edges toward 2.1%, which makes a single house macro assumption a planning defect.
September is when most calendar-year filers freeze the numbers that will govern next year, and this cycle the hardest choice is not the growth rate itself but how many growth rates the plan is allowed to have. The case for building 2027 planning assumptions around tariffs as two distinct macro cases rather than one blended house view rests on a finding hiding in plain sight in the latest CFO Survey: firms worried about trade policy are describing a measurably different economy than firms that are not. Averaging those two worlds into a single planning number does not reduce risk. It hides it.
The survey data now splits the economy in two
The CFO Survey, run by Duke's Fuqua School of Business with the Federal Reserve Banks of Richmond and Atlanta, points to a broad recovery in sentiment. Year-ahead real GDP expectations edged up to 2.1% in the Q1 2026 cut from 1.9% the prior quarter, and the average probability respondents assigned to negative year-ahead growth fell to 11.2% from 13.3%. The share of firms reporting that tariffs or trade-policy uncertainty affected their 2025-2026 price expectations or realizations dropped to 39% from 46.5% in Q3 2025. Spending excluding capex is loosening too, with 43% of firms reporting increases in the prior three months, up from 39.2%.
Underneath the headline, the cohorts diverge. In the Q3 2025 survey, tariff-concerned firms reported optimism of 59.9 against 64.3 for their peers and projected year-ahead real GDP growth of 1.6%, well below the figure for unconcerned firms. Tariffs and trade policy have ranked as the top respondent concern for three consecutive quarters, ahead of monetary policy and inflation, even as generic "uncertainty" fell from second place to seventh. Executives are no longer vaguely nervous. They have located the problem, and the location determines the forecast.
Why one house macro assumption is a planning defect
Most companies still hand finance business partners a single set of top-down assumptions: GDP, price, wage inflation, volume elasticity. That works when the shock is uniform. It fails when the shock is sectoral. A diversified industrial with a tariff-heavy components segment and an asset-light services segment that plans against one blended growth number is effectively setting the components team a target built on services conditions, and vice versa. The gap between roughly 1.6% and 2.1% sounds academic until it is compounded through volume, mix and a two-year capex commitment.
The practical fix is to split the planning assumption by trade exposure rather than by geography. Geography is the habit; exposure is the variable that actually drives the dispersion. That means tagging revenue and cost lines by whether they sit behind a tariffed input or a tariffed destination, then running the exposed set against a slower-growth, higher-input-cost case while the clean set runs on the base. Two cases, one plan, with the bridge between them documented so the board can see which assumption is doing the work.
Paid-but-litigated tariffs belong in the forecast, not the past
The second discipline is accounting for duties that have already been remitted but remain in dispute. EY's Technical Accounting Advisory group, in an interview published by the Forbes CFO newsletter in August 2026, addressed exactly this: how to handle forecasts, projections and investor communications when courts pause or overturn tariffs a company has already paid. The FP&A consequence is straightforward. If those payments are treated as sunk, a favorable ruling arrives as an unplanned windfall that distorts variance reporting, muddies segment performance and can quietly inflate incentive payouts against targets nobody set with a refund in mind.
Modeling the refund as an explicit recoverable branch, with a probability and a timing range, does two things. It keeps the windfall inside the plan's logic so variance analysis stays honest, and it forces a conversation with the audit and disclosure teams about what gets said publicly if the branch pays out. The same discipline EY applies to tariffs, they argue, now applies to AI spend: forecast a range, disclose the assumption, revisit it more often than annually.
Move the assumption review to quarterly and write down the trigger
CFO Dive framed the structural problem clearly in late 2025: unlike a demand shock or a rate cycle, trade policy keeps moving, so a forecast never gets the quiet period it needs to recalibrate once and settle. An annual assumption review is therefore mismatched to the cadence of the risk. The answer is not more scenarios but a shorter review clock with pre-agreed triggers, written down before the year starts. A named tariff rate change above a threshold, a court ruling on a specific duty, or a supplier surcharge above a set percentage should each automatically reopen the assumption set rather than waiting for the next planning cycle.
Guidance language is the last mile. If the external range a company publishes does not map to the same two cases the budget runs on, investors and sell-side analysts will build the reconciliation themselves, and their version will be less flattering. Aligning the disclosed range with the internal exposed and clean cases costs nothing in the planning process and buys credibility the first time a ruling moves the numbers.
Key takeaways
- Split 2027 planning assumptions by trade exposure, not geography, since exposure is what drives the dispersion in growth expectations.
- Tariff-concerned firms projected 1.6% year-ahead real GDP growth against 2.1% for the broad Q1 2026 CFO Survey sample.
- Model paid-but-litigated tariffs as a recoverable branch with probability and timing, so a refund does not distort variance reporting or comp.
- Replace the annual assumption review with a quarterly cadence and written triggers tied to specific rate changes or court rulings.
- Make the published guidance range map to the same exposed and clean cases the internal budget runs on.
