The river fell below zero. Two weeks later the barge quote doubled.

Why these three series belong together

A grain buyer in Tokyo, a barge line in St. Louis and a farmer in Iowa all meet in one number: the cost of moving a ton of corn down the Mississippi. That number is set by a quote, the quote is set by how much a tow can carry, and how much a tow can carry is set by the river stage. This edition puts the three public series side by side, dated, so that the chain can be read rather than assumed.

What the river did

On 1 July the St. Louis gage read 18.32 ft [F1] with a daily mean flow of 305,000 cubic feet per second [F4]. The stage fell through July and August and went below zero on 7 September; the low point was −1.22 ft on 9 September [F2], the day the flow bottomed at 90,300 cubic feet per second [F5]. October rains lifted the gage back to 10.18 ft by 17 October [F3], with flow at 194,000 cubic feet per second [F6].

What the barge quote did

The USDA weekly quote for downbound barges loading at St. Louis started the window at 15.52 USD per short ton on 7 July [F7]. It peaked at 834.7% of the benchmark tariff [F10], or 33.30 USD per short ton, on 22 September [F8], thirteen days after the river's low point. By 13 October it had eased to 26.87 USD [F9]: still 73.1% above the first quote of the window [F11], and 19.3% below the peak [F12].

What moved through Locks 27

Weekly grain through Locks 27 at Granite City peaked at 703,800 short tons in the week ending 25 July [F13], then collapsed to 84,300 short tons in the week ending 3 October [F14], three weeks after the stage minimum and eleven days after the quote peak. The last week of the window, ending 17 October, recovered to 349,600 short tons [F15]. Over the sixteen weeks the lock passed 6,355,344 short tons of corn, soybeans, wheat and other grain [F16].

The mechanism, and what else could explain it

Hypothesis · The sequence is consistent with a physical constraint: as the stage falls, the Coast Guard and the industry cut the allowed draft and the number of barges per tow, each tow carries less, operators need more tows per ton, and the quote per ton rises. The lags observed here, river low on 9 September, quote peak on 22 September, throughput low in the week of 3 October, are the order that mechanism predicts.

Hypothesis · Two other forces run through the same weeks. Late September is the seasonal trough before the corn and soybean harvest, so part of the throughput dip is calendar, not river. And harvest itself raises barge demand in October, which can hold quotes up even as the river recovers. The hypotheses below are written so that the next eight weeks can separate these explanations.

Who sits on this chain

On the river: ARTCo, ADM's barge line, earns the quote and loses tons per barge at the same time; CGB and Bunge originate Midwest grain and carry the freight in their bids to farmers. Each has public evidence of the activity, not of the volume on this exact route.

Hypothesis · Far from the river: Marubeni, Wilmar and Louis Dreyfus buy or move US Gulf grain and can switch origins; United Grain, Mitsui's Pacific Northwest gateway, is the alternative route that gains when the Gulf route is restricted; Agrotools serves Brazilian exporters who compete for the same Asian demand. These are deduced positions, labelled as such, to be confirmed or dropped as evidence accrues.

A conditional sensitivity, not a forecast

Scenario · If the quote rose a further 10% from its last value of 26.87 USD per short ton [F9], the increment would be 2.69 USD per short ton; on an assumed 10,000-short-ton cargo fully exposed to the quote, 26,870 USD. Contract terms, booking dates and origin substitution decide how much of that any single shipper actually pays.

What would change this reading

Hypothesis · A second drop below zero without a quote response would weaken the physical story; a quote that stays above 700% while the gage sits above 10 ft would point to harvest demand rather than draft limits; throughput that fails to exceed 600,000 short tons on an open river would mean the dip was not the river at all. Each of these is an open hypothesis with a date.