arcanum analytica
← field notes
balance sheets · monthly · usda + arcanum

balance sheets

us corn and soybean supply & disappearance — five marketing years of history, the latest two USDA WASDE prints, and the Arcanum Analytica end-of-marketing-year forecast with thesis prose. each commodity table is paired with a stocks/use → old-vs-new-crop spread regression chart.

release apr 2026 wasde · prior mar 2026

arcanum forecast column · Δ vs latest USDA · ember tint marks the house view and the stocks/use row

corn · 2025/26 marketing year
  2020/21 2021/22 2022/23 2023/24 2024/25 mar 2026 apr 2026 wasde arcanum Δ vs current
area planted (mil ac) 90.4 92.9 88.2 94.6 90.9 98.8 98.8 98.8
area harvested (mil ac) 82.2 85.0 78.7 86.5 83.0 91.3 91.3 91.3
yield (bu/ac) 171.4 176.7 173.4 177.3 179.3 186.5 186.5 186.5
beginning stocks 2,004 1,235 1,377 1,360 1,763 1,551 1,551 1,551
production 14,087 15,018 13,651 15,341 14,892 17,021 17,021 17,021
imports 24 24 39 28 22 25 25 25
total supply 16,115 16,277 15,066 16,729 16,677 18,597 18,597 18,597
feed and residual 5,667 5,671 5,486 5,831 5,454 6,200 6,200 6,100 -100
food, seed & industrial 6,466 6,757 6,558 6,880 6,814 6,970 6,970 6,990 +20
ethanol & by-products 5,028 5,320 5,176 5,489 5,436 5,600 5,600 5,620 +20
exports 2,747 2,472 1,662 2,255 2,858 3,300 3,300 3,380 +80
total use 14,881 14,900 13,706 14,966 15,126 16,470 16,470 16,470
ending stocks 1,235 1,377 1,360 1,763 1,551 2,127 2,127 2,127
stocks/use % 8.3% 9.2% 9.9% 11.8% 10.3% 12.9% 12.9% 12.9% +0.0
implied n/z spread (¢) +91.0 +50.4 +66.5 -23.5 +17.6 -5.9 -5.9 -5.9

n/z spread · current market quote · -26.00¢ · reciprocal fit · y = a + b/x · a = -146.91, b = +1821.48 · LOO ±21.6¢ · R²=0.823 · calibration: 2010/11 – 2024/25 (15 MYs)

6%8%10%12%14%16%18%-50¢+50¢+100¢+150¢stocks/use %n/z spread (¢)mkt -26.00¢10/1111/1212/1313/1414/1515/1616/1717/1818/1919/2020/2121/2222/2323/2424/25usdaarcanum
arcanum thesis · corn

house view leans modestly tighter than usda on the demand side. fas weekly sales are 3-4% ahead of the seasonal pace implied by the 3.30 bn export figure, and fgis shipments confirm the physical pull-through. corn-for-ethanol grind is sitting in the upper third of its 5-yr seasonal band, so the 5.6 bn ethanol assumption looks roughly right but with positive skew. feed offsets — placements light, cof inventory inline. net: ending stocks largely unchanged on the april print, but the direction of revision risk is lower stocks not higher into the july/aug WASDE. n/z model sits at modest carry given the comfortable s/u, and the live market is trading deeper carry still — if our tighter-demand call is right, the spread has room to roll back toward the regression line.

soy · 2025/26 marketing year
  2020/21 2021/22 2022/23 2023/24 2024/25 mar 2026 apr 2026 wasde arcanum Δ vs current
area planted (mil ac) 83.4 87.2 87.5 83.6 87.3 81.2 81.2 81.2
area harvested (mil ac) 82.6 86.3 86.2 82.3 86.2 80.4 80.4 80.4
yield (bu/ac) 51.0 51.7 49.6 50.6 50.7 53.0 53.0 53.0
beginning stocks 539 257 274 264 342 325 325 325
production 4,216 4,464 4,270 4,162 4,374 4,262 4,262 4,262
imports 20 16 25 21 29 25 25 25
total supply 4,775 4,737 4,569 4,447 4,746 4,612 4,612 4,612
crush 2,141 2,204 2,212 2,285 2,445 2,575 2,610 2,625 +15
exports 2,266 2,152 1,979 1,700 1,882 1,575 1,540 1,485 -55
seed 78 77 72 75 70 73 73 73
residual 33 30 42 44 23 39 39 39
total use 4,518 4,463 4,305 4,105 4,421 4,262 4,262 4,222 -40
ending stocks 257 274 264 342 325 350 350 390 +40
stocks/use % 5.7% 6.1% 6.1% 8.3% 7.3% 8.2% 8.2% 9.2% +1.0
implied n/x spread (¢) +163.6 +163.6 +155.5 +12.5 +18.3 +34.6 +34.6 +23.1 -11.5

n/x spread · current market quote · +6.25¢ · reciprocal fit · y = a + b/x · a = -69.12, b = +851.66 · LOO ±60.9¢ · R²=0.560 · calibration: 2010/11 – 2024/25 (15 MYs)

2.5%7.5%12.5%17.5%22.5%27.5%-100¢+100¢+200¢+300¢stocks/use %n/x spread (¢)mkt +6.25¢10/1111/1212/1313/1414/1515/1616/1717/1818/1919/2020/2121/2222/2323/2424/25usdaarcanum
arcanum thesis · soy

house view leans BEARISH ending stocks vs april usda — meaning we expect ending stocks to revise HIGHER, not lower, into the summer prints. the variance source is exports: fas weekly soy sales are running ~5% below the pace implied by the 1.54 bn export number, and brazilian fob is still trading premium-ish enough to keep marginal chinese tons going to santos rather than nola. crush is the offset — nopa march was strong on renewable-diesel pull, and oil stocks are tighter than seasonal, which biases the crush figure higher. the question is whether crush absorbs the export miss in full. our base case: it doesn't, so end-of-MY stocks finish ~30 mn bu above usda's april number. n/x model implies modest carry consistent with our looser stocks forecast; live market trading slightly tighter than our model suggests, watching for the spread to roll toward the regression line if exports continue to miss.

data: usda ers feed grains yearbook + oil crops yearbook · monthly USDA WASDE-670 (transcribed) · arcanum analytica forecast rebuilt 1 jun 2026 · curated 24 apr
esc to close · ↵ to jump · ↑↓ to select