Morgan Soy Simulator
Assessing the competitive impacts of changing soybean market fundamentals.
DNS Commodities · Confidential
You're making soybean capital and trading decisions using historical basis and historical draw territories. But the position you take this year has to pay off in a market that doesn't exist yet. Morgan builds that market. Add crush capacity, shock crop yields, shift export programs, reprice freight — and watch the competitive map redraw itself around your position.
Add capacity to a central Iowa crusher. Shift a quarter of Gulf export volume to the PNW. Lose a crop in the Eastern Corn Belt. In these imagined scenarios a basis history table can't explain the full competitive impact on your business. It can only inform an educated guess.
Morgan replaces that guess.
Build out your view of future crop year fundamentals and Morgan simulates it, showing soybean flows and relative values across the crop year for the entire US marketplace.
See for yourself.
Request access to see Morgan at work. Here's what's inside.
The baseline surface is Morgan's simulated output of the 24/25 crop year. The scenarios show what happens to the baseline when different market fundamentals change. These scenarios are illustrations of the type of custom work Morgan can do.
The national map for the 24/25 baseline crop year. Every county and every facility, valued month by month across the crop year, with a relative value referenced to either ADM Decatur or the US Gulf.
133k bushels per day of new capacity added in central Iowa, with the crop year re-simulated around it. Crop year carryout reduced to accommodate additional crush volumes.
Twenty-five percent of Gulf export demand redirected to the Pacific Northwest in the same shipment week. National export volume unchanged; nothing else adjusted.
A production shortfall in the Eastern Corn Belt, with the reduced supply pool asked to serve unchanged crush and export demand. A supply shock rather than a demand shock, a test of Morgan in a short year.
Site a plant, size an expansion, and see your delivered bean cost against every competitor in the country — before you commit the capital.
Understand how trade-flow shifts redraw draw basins, what moves from river to rail, and where the freight advantage sits week to week.
See where value genuinely sits in your geography, and what a structural change upstream or downstream does to your local market.
Independent diligence on a plant, a terminal, or an acquisition — grounded in fundamentals rather than seller's projections.
25 years of Ag commodity experience spanning
DNS Commodities provides fundamental insights to inform capital, operational, and trading decisions in your Ag/Renewable commodity business.
Most people meet Morgan and file it as a map product. It is much more than a map. It is a way of reasoning about where soybean value comes from, and the maps are how that reasoning gets displayed. These questions are written to make that distinction concrete.
Morgan is a simulation of the U.S. soybean market as a physical distribution problem. Every week of a crop year, Morgan coordinates movement of available bushels in each US county and sends them to their best demand point using a full freight matrix covering all major transportation modes.
Once movements are allocated across the crop year, Morgan uses its freight matrix to derive relative values for each county and each demand point for each week of the year.
In Morgan's framework value is a consequence of distribution economics and physical constraints.
No. Morgan will not tell you where November beans are going. It has no view on the board, on Chinese demand, on the Brazilian crop, or on next week's basis at your elevator.
No price of any kind is fed into the engine — not a futures price, not a posted bid, not a historical basis. If it did take a price in, the output would just be that price rearranged.
What Morgan produces is relative value: what a bushel in this county is worth compared to a bushel at Decatur, or at the Gulf, in this week, given where the demand sits and what freight costs.
With Morgan, it is possible to assign fixed relative values to supply or demand components and model distribution. For example, a client may ask what happens to export distribution if PNW values are locked at a $1/bushel premium over Gulf values. Morgan could then assign total export demand between the two regions to assess how many bushels each port is able to source while maintaining that relative value.
It is a framework — a structured way of asking "if this changes, what else has to change?" and getting an answer that is internally consistent across the whole country.
That consistency is the thing you cannot get any other way. You already know, from experience, that a new plant pulls beans away from its neighbours. What you cannot do in your head is trace that pull through 3,100 counties, across three freight modes, over 52 weeks, while every other plant and terminal in the country reacts at the same time. That is a distribution problem with hundreds of thousands of moving parts, and it is exactly the kind of problem a simulation exists to solve.
So Morgan is not a source of facts. It is an instrument for reasoning about consequences. Its output is a set of relationships you can interrogate.
A bid sheet tells you what. Morgan tells you why, and therefore what if.
Posted bids are the market's answer today. They are observations. They cannot be re-run under different conditions, they do not exist for the plant that has not been built.
Morgan is built from the causes rather than the symptoms, which means you can change a cause and watch the symptoms move. That is the only way to answer a question about something that has not happened yet.
Where the beans to feed it will likely come from, and what that does to everyone already competing for them. Morgan will also show the estimated impact on local county relative values as well.
133k bushels a day of new capacity at Des Moines. Morgan re-runs the crop year with the plant in place and reports: how far the draw territory has to reach, which counties enter the draw that were shipping nowhere near it before, and how the competitive bean cost ranking reorders.
Morgan reveals second order answers that are material to competitors and suppliers of the new capacity. For example it will show which existing plant loses its advantage because its marginal counties got bid away, not to Des Moines, but to a different crusher impacted by Des Moines' new demand.
Trade flows shift for reasons nobody in the supply chain controls. The question is always the same: when the destination changes, which origins get pulled and which get stranded.
Twenty-five percent of Gulf export demand redirected to the Pacific Northwest, in the same shipment week, with national export volume unchanged. Nothing else touched.
Morgan reports how far east the PNW draw basin has to stretch to fill the new program, what comes off the river as a result, which railroads pick up the freight, and what it does to the delivered bean cost of plants that were never part of the export trade at all.
Morgan is well suited to tackle changes in freight economics or service.
Every Class 1 carrier is distinct, with its own service network, rate structure, and market access. Morgan carries this structure — loading points, demand points, mileage, and rates. Utilizing public rate tariffs where available and estimated rates where necessary, Morgan reveals how shifts in rates, freight availability, supply or demand impact the soybean market.
So you can ask: if this corridor's economics move, which counties lose their cheapest route to demand, where do those bushels go instead and what does the substitution cost?
If you are a railroad, the same question runs the other direction: which bushels are you positioned to win, and what would it take.
A short crop is not a supply problem in one place. It is a redistribution problem everywhere, because demand that used to be served locally has to go looking.
A 30% yield loss across Ohio, Indiana, Michigan and Kentucky, 15% in Illinois, with crush demand left unchanged. Roughly 342 million bushels off national production.
The question Morgan is being pointed at: how far west do eastern draw territories have to stretch to cover the gap, which eastern plants lose the freight advantage they normally enjoy, and what does the additional length do to the river and rail networks in between.
It can tell you how a plant in a given place would compete for beans, how far it would have to reach to fill itself, who it would be taking from, and how that changes across the seasons of a crop year. Run several candidate sites and you have a defensible ranking.
Plant siting turns on rail access, utilities, permitting, labor, meal and oil markets and the price of the land — none of which Morgan models. It answers the origination question, which is usually the one with the most money attached and the least rigor applied. Later iterations of Morgan will address the meal and oil markets using a similar analysis to soybeans. Stay tuned!
A lot goes into a daily price for soybeans at a facility. Futures price moves, carry vs inverted futures structure, corn crop size, plant downtime, a local weather event, missed RR service, a bin quality issue — all impact prices every day. Morgan does not try to capture any of that information and therefore is not good at some things.
The three scenarios on this site are the demonstrations, not the menu. They were chosen because they are easy to explain, not because they are the limit. What a custom engagement looks like depends entirely on who is asking:
In every case the pattern is the same: your data goes in, your question gets framed as a scenario, and the engine reports the structural consequence.
Yes. Client data is used for that client's work and is not folded into the public build or into anyone else's scenario. Custom scenario results belong to the client who commissioned them. NDAs for all work are expected and a normal course of business for DNS Commodities.
Request access to this site. Tell us who you are and what you're working on, and we'll set up your login so you can see the baseline surface and the scenarios for yourself. When you find a question Morgan should be pointed at, commission a scenario from inside.
Assessing the competitive impacts of changing soybean market fundamentals.
DNS Commodities · Confidential
A production shortfall in the Eastern Corn Belt, with the reduced supply pool asked to serve unchanged crush and export demand. A supply shock rather than a demand shock, a test of Morgan in a short year.
Locked before the build, as with every scenario on this site.
| Input | Change |
|---|---|
| Yield — Ohio, Indiana, Michigan, Kentucky | −30%, every county |
| Yield — Illinois | −15%, every county |
| Crush demand | Unchanged, everywhere |
| Export demand | Reduced pro-rata across all export sinks |
| Carry-in | Unchanged |
| Carryout | Reduced, with every quarter-end set rather than left to the model |
What this scenario will report when the results publish.
| How far west eastern draw territories stretch to cover the gap |
| Which eastern plants lose the freight advantage they normally enjoy |
| What the additional length does to river and rail loads in between |
| Which counties gain relative value, and which give it up |
The three scenarios on this site are examples of the type of work Morgan can do. Tell us the question in front of you and we will reach out within 2 days to schedule a conversation to discuss scope, timeline, and cost.
Call DNS Commodities directly, or write to duane@dnscommodities.com.
+1-662-550-9462
In this example we model a 133k bushel/day crush expansion in Des Moines, IA. Morgan re-simulates the 2024/25 crop year with the increased capacity in place and reveals how the market structure changes: where the beans come from, how far they travel, what they cost, and how the competitive reorder shifts.
No price was assumed. Morgan updates demand with the crush increase and models the impact. Outputs are derived from Morgan's simulation engine that importantly does not set a soybean price level. Morgan's value outputs emerge from county supply, plant/export demand and Morgan's freight matrix.
How supply responds. The added crush is funded from carryout rather than from the historical selling pace. Each state's quarterly disappearance schedule is raised so that 53.5 million bushels more of the same supply pool clears over the crop year — 13.375 million per quarter, allocated across states in proportion to each state's disappearance in that quarter. Production is unchanged; ending stocks fall by the amount the plant consumes.
Baseline is the simulated 2024/25 crop year. Scenario is the same year re-run with 133k bu/day of additional Des Moines capacity. All other supply, crush, export and freight matrix data is left unchanged. In this scenario, annual carryout is reduced by the modeled crush increase.
| Measure | Baseline | With expansion | Change | Note |
|---|
Counties supplying the plant, and how much each ships across the crop year. The territory grows from — counties to —.
Change in volume drawn, by county. — counties enter the draw from zero; — existing suppliers ship materially more.
Bean cost relative to ADM Decatur, a central reference point. A higher number means the plant must pay more for a delivered bushel. Changes inside two cents are within the model's estimation band and are shown in grey.
| Plant | Baseline | With expansion | Change | Rank before | Rank after |
|---|
Change in county bean value. — counties gain a cent or more, averaging —. This output is useful not just for the owners of the Des Moines plant, but also other stakeholders impacted by increased soybean prices in the region.
Want this run on your own plant, your own geography, your own assumptions?
Twenty-five percent of every Gulf export sink — Mississippi River, East Gulf, North Texas and South Texas — is redirected to the Pacific Northwest in the same shipment week it originally moved. The added volume is split between Columbia River and Puget Sound in proportion to their existing share.
Nothing else changes. National export volume is identical. Crush demand, production, and every state's quarterly stocks schedule are left exactly as the baseline. The model was given no price, and no supply assumption was adjusted to accommodate the shift.
The question it answers. If trade flows move west, which basins gain, which railroads carry the freight, what happens to the river, and what does it do to the delivered bean cost of plants already in the ground.
Baseline is the simulated 2024/25 crop year. Scenario is the same year with a quarter of Gulf export demand relocated to the PNW.
| Measure | Baseline | With shift | Change | Note |
|---|
Counties supplying Columbia River and Puget Sound, and how much each ships across the crop year. The basin grows from — counties to —.
Change in volume drawn, by county. — counties enter the basin from zero; — existing suppliers ship materially more.
Freight rates are identical in both runs. The same freight matrix, the same published rail tariffs and fuel surcharges, the same barge rates, the same diesel, the same river conditions. Nothing in the cost of transportation was altered. Every difference below comes from beans taking different routes to different destinations — not from a change in what it costs to move them.
Single-line movements only. Interchange traffic between two railroads is excluded so each carrier is credited with the haul it actually originates and terminates.
| Carrier | Baseline MBu | With shift MBu | Change | Baseline ton-miles (MM) | With shift ton-miles (MM) | Avg haul before | Avg haul after |
|---|
Volume and ton-miles moving by river, by month.
| Month | Baseline MBu | With shift MBu | Change | Baseline ton-miles (MM) | With shift ton-miles (MM) | Avg miles before | Avg miles after |
|---|
County bean value in each month of the crop year, baseline beside scenario. Values are spreads, not price levels — the model sets no price. Colour scales are fixed across every month so the maps are comparable as the slider moves.
Delivered bean cost relative to ADM Decatur, crop-year average. A higher number means the plant must pay more for a bushel at its gate. These plants are not running more volume — they are competing with a Pacific Northwest that is bidding for the same crop. Changes inside two cents are within the model's estimation band and shown in grey.
| Facility | Baseline | With shift | Change | MBu baseline | MBu scenario |
|---|
Corridor value relative to ADM Decatur. The Pacific Northwest must bid up to source the additional volume; the Gulf, buying less, eases.
| Corridor | Baseline | With shift | Change | MBu baseline | MBu scenario |
|---|
Want this run on your own program, your own corridors, your own assumptions?
This baseline value surface displays Morgan's simulated results using data from the 2024/25 crop year. Across the top is a month slider, allowing a look at each month's values at the County or Facility (crush and export ports) level. Toggle between these views on the left side. Additionally, filter the output to Crush plants, Export ports, or All. When in Facility view, select a plant and see a chart of its weekly values for the crop year below the map.All values are expressed either as relative to ADM Decatur IL or US Gulf, both major price reference locations in the trade.