Section 1
The framework: four models on a moving regime
Model one: the timing mismatch as the core defect Name the defect precisely, because the name contains the fix. The problem is not that materials cost more. Prices rise and fall in every era, and a competent operator prices to the level. The problem is the timing mismatch: you commit to a fixed delivered price at bid, and you buy the steel weeks or months later, and in that interval a single proclamation can move the commodity 20 to 50 percent under a number you are already bound to. The loss lives in the gap between the two events, not in the price level itself. Widen the gap and you widen the exposure. Shrink the gap and you shrink it, regardless of where prices go. This reframing is the whole article in one line: the interval is the risk, so the interval is the lever. Put numbers on it. Take a job with 30 percent of contract value in structural steel and a bid that stays live for 60 days before you purchase. If a proclamation lifts steel 17 percent inside that window, the move on the total contract is roughly 5 percent of contract value, which on a thin service-business margin can erase the profit on the job outright. Now cut the live window to 10 days and qualify the buy the week you sign. The same 17 percent move can only reach you if it lands in a far narrower slice of time, and the probability of a jump inside 10 days is a fraction of the probability inside 60. Nothing about your forecast changed. Only the interval did, and the interval is the thing you control. Timing-gap model: the core lens • Assumes: the loss is a function of the interval between price commitment and material purchase, not of the price level. • Fits because: a fixed price only hurts you if the input moves after you lock it and before you buy it. • Breaks when: the input is bought effectively at bid, so the gap is near zero and the model has nothing to bite on. Also breaks if the loss is driven by a level you mispriced rather than a move you were exposed to. • Counteracts: the urge to fix a timing problem with a forecasting tool. • May reinforce: an overfocus on speed that ignores the times when waiting is correct. Model two: the proclamation as a jump process, not a diffusion Ordinary price risk is a diffusion: prices wander up and down in small steps, and volatility is roughly continuous. Forecasting and hedging tools are built for that world. A tariff proclamation is a jump process: the price sits flat, then gaps hard on an announcement, then sits flat again. The two require opposite responses. Against diffusion you can hold a position and ride the noise, because small moves average out. Against jumps you cannot, because a single gap can exceed a year of ordinary volatility in one night, and it does not average out. Once you see 2025 as a jump regime, the case for prediction collapses and the case for structural protection, clauses and short windows, becomes obvious. You do not forecast a jump. You build so a jump cannot reach a price you have frozen. Jump-process model: the volatility lens • Assumes: the dangerous moves arrive as discrete steps on announcements, not as continuous drift. • Fits because: tariff changes are events at a minute, not trends over a quarter. • Breaks when: the regime is quiet and prices genuinely do diffuse, in which case ordinary trend pricing is fine and heavy structural protection is overkill. • Counteracts: the assumption that recent price stability means the next move will also be small. • May reinforce: perpetual defensiveness, carrying the cost of protection through calm periods that did not need it. Model three: dynamic sourcing as a real option Flexibility in where and when you buy is an option, and options have value precisely when the future is uncertain. A contractor locked to a single supplier and a single origin has sold that option away. A contractor who can switch suppliers, substitute a domestic mill for an imported one, redesign a spec to a less-exposed material, or move a purchase earlier or later, holds a portfolio of options that pay off exactly when a proclamation reshuffles which source is cheapest. Because most derivatives were tariffed on their steel or aluminum content and not on domestic-sourced metal (Bureau of Industry and Security, on the 2025 derivative treatment), the source and the spec became live levers, not fixed givens. Dynamic sourcing is not just procurement hygiene. It is the operational form of holding options in a jump regime. Real options via sourcing flexibility: the adaptive lens • Assumes: you can value and hold the right to switch source, spec, or timing, and that flexibility pays off under uncertainty. • Fits because: a proclamation changes which supplier and which origin is cheapest, and switching captures that. • Breaks when: switching costs are prohibitive, specs are locked by the owner or code, or lead times make substitution impossible. A locked spec on a long-lead item has no optionality to exercise. • Counteracts: single-source complacency built in calmer years. • May reinforce: churn, chasing marginal savings across suppliers at a coordination cost that exceeds the gain. Model four: comparative statics to size the move The plain first-order check still earns its place. Move the tariff rate, hold everything else, trace the effect: rate up means delivered cost up means margin down for whoever holds the frozen price. It gives you the sign and a rough size before you act, and it keeps you from over-engineering a response to a small exposure. But it carries the same warning as always, and here that warning is the whole point. Comparative statics: the first-order lens • Assumes: ceteris paribus, one variable moves against a stable structure. • Fits because: you need the direction and rough magnitude of a tariff move before choosing a response. • Breaks when: the shock is large enough to change the structure, which is exactly what a proclamation does. Then the stable curve it traces no longer exists. • Counteracts: overreaction to a trivial exposure. • May reinforce: false confidence in a relationship that a regime break has already voided.
Section 2
The structure break: why forecasting is the wrong tool, stated plainly
This is the flag the whole article turns on. Every forecasting method, formal or intuitive, assumes the process generating tomorrow's price looks like the process that generated the last several years. It fits a relationship on history and projects it forward. A tariff-by-proclamation economy breaks that assumption on purpose and on a schedule nobody controls. The move does not emerge from the supply-demand data your model reads. It arrives from a signature at 12:01 am, and the signature is not in the training data. Fit a beautiful model of steel prices on 2015 through 2024 and it will tell you nothing about June 4, 2025, because June 4 was a policy event, not a market event. That is why the answer is operational rather than predictive. When the regime can break by announcement, you stop trying to see the break coming and you build so the break cannot reach a price you have committed. You shorten the interval, you keep sourcing flexible, and you default to clauses that share the risk. Every one of those is an operating decision you control, and none of them requires you to predict a proclamation you cannot predict. Down-weight the forecast. Foreground the portfolio and the base rate.
Section 3
The solution: GEER, then RADAR, then CHAIN
GEER: rank the operational levers Size the exposure first: Proclamation exposure = tariff-exposed share of contract value × length of the interval between price lock and material purchase × your inability to switch source, spec, or timing. The interval and the switching constraint are the two terms you actually control. Pull the levers cheapest and most reversible first: 1. Shorten the bid validity window. Quote good for 10 to 14 days rather than 30 to 60. This directly shrinks the interval term, costs nothing, and is fully reversible. It is the highest-leverage operational move available. 2. Make an escalation clause the default. Attach a ConsensusDocs 200.1 or equivalent index-based clause to every bid so risk-sharing is the starting point and removing it is the other side's job. Covered in depth in the companion piece on the escalation clause divide. 3. Match supplier quote validity to bid validity. Align the two windows so no risk falls into the gap between your locked bid and your expiring supplier price. 4. Build sourcing optionality before you need it. Qualify a second supplier, identify a domestic substitute, and know which specs can flex to a less-exposed material. Do this in calm, so the option exists when a jump hits. 5. Pre-buy the single highest-volatility long-lead item. Exercise the option to lock only where a gap is most likely and most damaging. Flag this one as capital-required, since it ties up cash and storage. Notice what is not on this list: a subscription to a better price forecast. It would sit below every operational lever, because it optimizes the one variable you cannot control. RADAR: the dated portfolio Do now (zero-regret): shorten every bid validity window, attach the escalation clause as default, and itemize tariff-exposed content per bid. These help under a jump and cost nothing under calm. Hedge (cheap, bounded): qualify a backup supplier and a domestic substitute for your most exposed input, add a named tariff allowance line, and carry a small buffer of only the single most volatile material. Each caps a large loss for a modest, bounded cost. Defer with a trigger (irreversible, pre-commit the response): the large pre-buy, the warehouse lease, a full sourcing re-architecture. Do not do these on a hunch in a jumpy week. Write the trigger now: if a Section 232 proclamation covering your inputs is posted, or the steel or aluminum PPI moves past an agreed monthly threshold, you execute the pre-buy on a named list and switch to the pre-qualified backup source. The discipline is deciding the response before the event, so the event finds a plan instead of a panic. CHAIN: the base-rate check Place 2025 in its reference class of material-shock episodes: 2021, 2018, 2008. The base rate is stable across them. Operators who had shortened their exposure interval and held sourcing flexibility rode the shock with intact margin. Operators who had frozen long-dated fixed prices against a single locked source absorbed the move. The reported figure that roughly 70 percent of contractors saw margin compression on pre-tariff fixed-price contracts fits that history (industry survey reporting; treat the precise percentage as a single-source estimate rather than a hardened statistic). History rewards structure over prediction. The matrix-break flag: proclamation-driven tariffs shorten the interval between shocks, so the base rate of "a jump lands inside my exposure window" is rising above what the historical record shows. The frequency itself is trending up. That is the case for making short windows and default clauses your permanent operating posture, not a temporary reaction that relaxes the moment the news quiets down.
Section 4
What the framework cannot see
This ensemble tells you how to structure pricing and sourcing so an unpredicted proclamation cannot reach a frozen price. It is blind to three things, and you should say so. It cannot see demand. Every lever here protects your margin on work you win. None of them tells you whether the work exists. If tariffs lift total project cost enough that owners cancel, defer, or shrink their pipelines, your short windows and flexible sourcing protect a smaller and smaller book. The best operating tempo in the world cannot manufacture demand that the same tariffs destroyed. It cannot see competitive erosion. If every shop in your market shortens windows and defaults to escalation clauses, the collective advantage disappears and you are back to competing on price with the added friction of shorter quotes. Structure is a durable edge only while others lack it. The framework prices your risk. It does not guarantee the risk premium survives everyone adopting the same defense. It cannot see the operational cost of agility itself. Qualifying backup suppliers, maintaining substitute specs, and re-quoting on short cycles all consume administrative time and coordination that a small shop may not have. The models value flexibility as if holding it were free. It is not. An operator can theoretically hold every option and practically drown in the overhead of managing them.
Section 5
The fitness test
You are ready to trade forecasting for operating, and you should make the switch now, if these hold. You can measure the interval between price lock and material purchase on your jobs and you have the standing to shorten your bid validity window without losing the work. You can attach a default escalation clause and qualify at least one alternative source or substitute spec. And your book is large enough that protecting margin per job matters more than the overhead of running the defense. If that is you, stop buying forecasts of a signature you cannot predict and start shrinking the interval you actually control. You should stay with simple level-pricing, and skip the machinery, if the opposite is true. Your jobs are short-cycle and you buy materials within days of the bid, so the interval a proclamation could exploit is already near zero. Your tariff-exposed share is low enough that the exposure does not justify the overhead. Or you are a small enough shop that the administrative cost of running backup suppliers and short-cycle re-quotes would cost you more than the risk it hedges. In that case, price the risk into the number honestly, keep your sourcing simple, and revisit the moment your interval or your exposure grows. The tools are a response to a specific exposure, not a virtue to adopt for its own sake.