WSCFI dips 2.06% to 3,420.68 as Persian Gulf and Red Sea risk keeps Gulf lanes soft

The Analysis

On 25 September the Wuhan Containerized Freight Index closed at 3,420.68, down 2.06 per cent, with demand described as stabilising. Strictly speaking, that number describes no shipment anyone reading this actually booked. The composite is an average of lanes that moved in opposite directions, and the limbs are the interesting part: Persian Gulf and Red Sea down 4.33 per cent, Europe down 1.24, the Mediterranean down 2.65, the US East Coast up 0.54, the US West Coast down 1.05, Southeast Asia and Busan up between 3.74 and 4.79.

A contract that indexes to that composite has therefore agreed a number that neither party ships. That is the question this piece takes up, and it is a question about drafting and evidence rather than about the direction of freight. The index is not wrong. It is being asked to do something it was never built to do.

What follows reads the divergence before the average, separates a rate movement from a hazard as legal categories, sets out where index-linked clauses usually fail, works two examples with the assumptions stated, then takes the timescale apart and names the one condition that would make me put the whole argument down.

An index is a statistic before it is a price. It has a basket, weights, a survey method, a publication calendar and a revision practice, and every one of those is a choice made by whoever compiles it. On 25 September the composite closed at 3,420.68, down 2.06 per cent on the previous reading, against a background of demand that the record describes as stabilising. That is a description of a market, and a description is all it is until a clause adopts it. The clause, not the index, does the work of turning a statistic into an obligation, which is why two contracts referencing the same number can carry entirely different risk.

The limbs of the average are worth reading before the average is. Persian Gulf and Red Sea lanes fell 4.33 per cent with risk still elevated. Europe fell 1.24 per cent and the Mediterranean 2.65 per cent on weak demand. The US East Coast rose 0.54 per cent while the US West Coast fell 1.05 per cent, the two coasts pulling apart on capacity rather than on demand. Southeast Asia and Busan rose between 3.74 and 4.79 per cent. Set the highest limb against the lowest and the spread is a little over nine points, around a composite that moved two. An average of that composition is arithmetically correct and, for any single lane, uninformative. It does not even bound the outcomes, since it tells you nothing about the highest or the lowest figure inside it.

The composition of the basket explains how the average can move so little while the parts move so much. Weights in a composite are not fixed in practice: they follow the volume actually carried on each lane, so a lane that loses capacity and a lane that gains it will drift in importance between publications without any reweighting being announced. The record attributes the divergence to demand on some lanes and to capacity on others, and the two coasts are the clearest case, one rising on firm demand while the other eases on ample capacity. That is equipment being reallocated between trades, and a statistic weighted by volume registers it as a small net move over a large internal one. Reading the composite as a summary of pressure means reading a number that has already averaged the pressure away.

A rate that falls while risk rises is not evidence of slack. On the Gulf and Red Sea lanes the two moved together, and the fall reads better as a price for a hazard than as a discount for weakness. The distinction is not rhetorical, because contracts treat the two as different animals. A market movement is an ordinary commercial risk, carried by whichever party the contract puts it on. A hazard is the raw material of force majeure, of war risk cover, and of the argument that performance has become something other than what was promised. The same 4.33 per cent is therefore two different facts depending on which question is being asked, and the index reports only one of them.

The legal distinction behind that is old and it gets lost in commercial conversation. Force majeure requires that performance be prevented, not that it become more expensive. A rate that doubles does not excuse a party from performing; it moves money between parties. That is why a rate movement almost never triggers a force majeure clause, and why parties who want protection from cost write a hardship clause, a price reopener, or an index. Here is the exception, and it is a real one: where a clause defines its trigger by reference to a named route, a named chokepoint or an official advisory rather than by reference to cost, the analysis changes completely, because the trigger is then an event and not a price. The Bab el-Mandeb and the Strait of Hormuz appear in contracts for exactly this reason. A clause that names them is testing whether something happened. A clause that names the index is testing what an average did.

The service being priced deserves its own sentence, because a cheaper quotation on a route with elevated risk is not the same service at a lower price. Carriage terms commonly give the carrier liberty to divert or to omit a call, and where that liberty is exercised the transit the shipper expected is not a contractual entitlement unless it was separately agreed with a consequence attached to it. The discount and the discretion therefore arrive together. The place to check is not the quotation but the terms behind it: a quotation states a number, and the terms state what the number actually buys.

Most index-linked clauses fail in one of five places, and the failures are drafting failures rather than market failures. They name the index but not the sub-index, so a composite is applied to a lane. They name the sub-index but not the publication and the vintage, so a later revision reopens an invoice that both sides had already settled. They set a threshold without a cure period, which makes a brief breach permanent, or a permanent breach invisible, depending on which side held the pen. They cap the adjustment in one direction, usually the direction that protects the drafter. And they assume the index will go on publishing, which is the assumption least likely to hold in precisely the weeks when the clause starts to matter.

The reset date deserves separate mention, because it is the one variable both parties can see and neither usually negotiates. A clause that resets quarterly applies a single reading taken on a fixed day to three months of cargo, and in a market that moves this way the choice of that day is worth more than the width of the collar. The vintage raises the same problem in a quieter form: where a compiler revises a period after publication, a settlement on the first print and a settlement on the revision are different settlements, and a clause that does not say which governs has handed the difference to whoever notices first. These are unglamorous points, and they decide more disputes than the headline rate does.

A worked example, with the assumptions stated because the contract terms are mine and not the market's. Assume a customer agreement that indexes the freight element to this composite, resets quarterly, and carries a collar of plus or minus 3 per cent, so that nothing happens inside the band. Assume a quarterly freight spend of USD 2 million, and that the cargo moves on a Southeast Asia lane which rose 4.79 per cent. The composite fell 2.06 per cent, which is inside the collar, so no adjustment is due. Your cost rose 4.79 per cent, which on that spend is USD 95,800 in the quarter, or USD 383,200 across a year of four such quarters. Nothing in the outcome is a breach. The clause did exactly what it says it does. The exposure sits in the gap between what the clause measures and what the cargo does.

The Gulf case runs the other way, and it is where disputes actually begin. Assume a quarterly spend of USD 1.5 million on a Persian Gulf lane whose rate fell 4.33 per cent. The customer reads the fall and asks for it to be passed through, and 4.33 per cent of that spend is USD 64,950 a quarter. The rate did fall, and that is not in dispute. What the index does not price is the rest of the same package: a diversion, an omitted call, the additional transit, and whatever the cover costs when the route in question is the one being avoided. The discount is real in the index and contingent in the result, and a clause that indexes only to the rate transfers the certain part while leaving the uncertain part exactly where it was.

The opposite case is worth stating, because it is the one that gets defended at the negotiating table. Where a lane rises and the reference is a composite, the customer's cost rises only if the composite clears the collar. Smoothing of that kind is a virtue for budgeting and a defect for allocation, and one clause produces both. A party that wants a predictable number should index to a composite. A party that wants an accurate allocation should index to the lane. Drafting a single clause to do both is how a contract ends up doing neither, and the failure only becomes visible in a quarter like this one, when the limbs separate by nine points.

Where no clause exists at all, the index does not vanish from the relationship, it changes function. It stops being a term and becomes an argument, and in a price conversation an argument is asymmetric by construction, because the party raising it chooses the period, the lane and the limb. Facing a customer who cites a 4.33 per cent fall, a supplier has three responses available and none of them is a mechanism: dispute the relevance of the lane, dispute the period, or accept and recover the difference somewhere else. That is not the equality a clause creates, and it is why the absence of an index clause gets read as neutrality by the side that benefits from the absence. Strictly, it is a default rule written by nobody and applied by whoever speaks first.

Evidence is the other half of the question, and it is where an index is weakest. In a dispute the tribunal asks what a party knew and when, what it was told, what it did about it, and whether its conduct was reasonable in the circumstances. A composite answers none of that. It says what the average of a basket did. The primary evidence is your own booking record, the quotations you received, the notices you were given about diversions or omissions, and the dates on which each of those arrived. Those documents exist in every company and are assembled in almost none, which is why the same dispute is cheap for one side and expensive for the other, and the difference is clerical rather than legal.

Notice is the part of the record that decides most of these cases and receives the least attention. A diversion or an omission of call is usually communicated, sometimes late and sometimes to the wrong address, and what a party does in the days after that communication often matters more than the event. Silence after notice is read in more than one way depending on the forum, and none of the readings favours the party that stayed quiet while the cargo went the long way round. The discipline is unremarkable and it is rare: keep the notice, date it, answer it, record what you did about it. None of that needs a lawyer at the time, and all of it is what a lawyer asks for two years later.

Scale changes the shape of the problem more than its size. A large importer running a portfolio of index-linked contracts has a weight-drift exposure it cannot see: as lanes with different rates move, the implicit weights inside the composite shift, and a clause that fixed only the name of the series has fixed nothing about the basket. A mid-size importer usually has one contract of this kind, with one important customer, and the question is personal, because raising the collar at renewal costs goodwill that is hard to price. A small importer frequently has no clause at all, which is its own asymmetry: a customer can cite a 4.33 per cent fall in a price conversation and decline to mention a 4.79 per cent rise, and without a clause there is no mechanism to appeal to, only an argument. The absence of a clause is not neutrality. It allocates the risk to the party with less room to refuse.

Read the lanes as a portfolio and the arithmetic turns awkward. Europe at minus 1.24 and the Mediterranean at minus 2.65 mean that on those lanes an index-linked clause reduces what you recover in a soft market, which is the clause working as designed and still unwelcome. The US East Coast at plus 0.54 against the US West Coast at minus 1.05 mean one reference can move in opposite directions on two lanes serving a single customer relationship. Southeast Asia and Busan at plus 3.74 to 4.79 mean the collar binds exactly where demand is strongest, which is where the cargo is least optional. A book spread across those four positions and indexed to one composite nets out close to 2.06 per cent. That looks like stability. It is the cancellation of four different exposures into a number that describes none of them.

The two coasts make the point in a way a composite cannot. One customer relationship can move cargo through an East Coast service that rose 0.54 per cent and a West Coast service that fell 1.05 per cent in the same week, against a clause that references neither. The allocation question then becomes internal: which lane carries the indexed element, and whether that answer was ever written down. In most companies it was not, so the answer defaults to whichever lane the finance team happened to model, which is a fact about a spreadsheet rather than about the contract. Where both coasts serve the same product, the divergence is 1.59 points between two services the customer experiences as one.

The timescale is quarterly, which is why the reading matters more than the level. A publication on 25 September feeds a reset in the fourth quarter, and a fourth-quarter reset feeds the annual tender, so a divergence observed now is still working through documents in the first quarter of next year. The useful lead times are short and specific. Four to six weeks before a reset is the window for testing whether the collar will bind on the lane you actually ship, and about three months before a tender is the window for deciding whether the reference series should change at all. Both are drafting windows rather than trading windows, and the second closes long before anyone feels the first.

The point the coverage does not make concerns dispersion rather than direction. A composite that moves 2.06 per cent while its limbs span a little over nine points is not a description of a market; it is a description of an average, and the dispersion is the part carrying the information. Compilers publish the composite. Contracts copy the composite. Nobody publishes the dispersion as a term and nobody drafts against it. Had the clause referenced the lane, or the sub-index, or even the composite together with a stated tolerance for divergence, the drafting cost would have been a line of text and the exposure would now be visible. The reason it is not done is not technical. The composite is the only number both sides recognise without an argument, so recognition is doing the work that accuracy is supposed to do, and the bill for that substitution arrives in quarters like this one.

Three counterfactuals bound how much of this is worth doing. Had the composite moved six per cent instead of 2.06, the three per cent collar would have been breached and the same divergence would have produced a payment in the opposite direction, so the outcome turns on the collar rather than on the market. Had the limbs stayed within a point of one another, the composite would have been an adequate proxy for every lane and this analysis would collapse into a sentence. And had the clause referenced the lane sub-index, the 2.06 per cent would be irrelevant to the contract, which is the cleanest test of whether the drafting or the market is doing the damage. The condition that would make me revise the reading is a narrowing of the spread, and a narrowing is the one thing a single week's publication cannot establish.

The 2.06 per cent is a public fact and it will be quoted in a hundred places this week. Whether it governs your contract is a private fact, settled in a clause drafted when nobody in the room was looking at a nine-point spread. That clause is still open at renewal, and the distance between a composite reference and a lane reference is one line of text and most of the exposure.

↑ Back to the key points

— By Dr. Ingrid Voss