THE FARMERS FIRST AG INDEX

FFAI v3.0 METHODOLOGY — dashed terms have plain-English tooltips, hover or tap

⇩ Download Full Methodology PDF

WHAT IT IS

The Farmers First Ag Index measures U.S. agricultural financial conditions using 16 publicly available Federal Reserve economic data series?Free government data published by the St. Louis Fed. Same numbers the banks and USDA use. Anyone can look them up at fred.stlouisfed.org.. It predicts the direction and magnitude of agricultural loan delinquency?The percentage of farm loans that are 30+ days late on payments across all U.S. banks. When this goes up, farmers are struggling to pay their bills. When it goes down, things are good. — the percentage of farm loans 30+ days past due across all U.S. commercial banks. When this number rises, farmers are under financial stress. The FFAI gives you that signal in real time, decomposed by sector.

THE COMPOSITE (Soy + Fed Funds)

Two inputs drive the validated core: soybean futures prices and the Federal Funds interest rate?The rate banks charge each other overnight. When the Fed raises this, your operating loan rate goes up too. It's the master dial for borrowing costs across the whole economy.. Soybeans proxy overall crop revenue conditions. The Fed Funds rate captures debt service costs?What it costs you to carry your loans — operating lines, equipment notes, land payments. When rates go up, every dollar you owe costs more to service. — agriculture is among the most debt-intensive industries in the U.S. economy, with production loans typically on variable rates. When both are favorable (high soy, low rates), farmers prosper. When both turn adverse, delinquency rises.

The model uses expanding-window regression?Imagine you're standing in 2015 making a prediction. You can only use data from 2003-2014 — you can't peek at 2016. Then in 2016, you add one more year and predict again. This proves the model works in real time, not just in hindsight.: at each quarter, it trains only on data available up to that point, preventing look-ahead bias?Cheating by using future information to make a "prediction." Like saying you predicted the 2012 drought after it already happened. Our model can't do this because it only sees past data at each step.. This is how it would have actually performed in real time.

VALIDATION

Tested against every quarter since 2003 (91 observations). Leave-one-out cross-validation?Take out one quarter, build the model with the other 90, then predict the one you removed. Do that 91 times — once for every quarter. If the model still works after 91 tests, it's not a fluke.: r = 0.49 (p < 0.000001)?The "p-value" is the chance this result is just random luck. p < 0.000001 means less than 1 in a million odds this is a coincidence. For context, most studies accept 1 in 20 (p < 0.05) as meaningful.. Expanding-window out-of-sample: r = 0.51 (p < 0.00001). Survives Bonferroni correction?When you test a lot of models, some will look good by pure chance. Bonferroni accounts for this by making the test much harder to pass. We tested ~50 models. Our result still passes after this penalty. It's real, not cherry-picked. for 50 multiple comparisons. 73% regime accuracy — when the index says above 50, delinquency is below median 73% of the time.

The model explains 28% of delinquency variance (R² = 0.28)?Think of it this way: about 28 cents of every dollar of farm loan trouble can be explained by crop prices and interest rates. The other 72 cents is weather, trade deals, your neighbor's management decisions, and everything else. 28% from just two numbers is actually strong.. The other 72% is weather, trade policy, individual farm management, regional conditions, and crop insurance decisions. The FFAI is a conditions indicator, not a crystal ball. We state this because intellectual honesty matters more than marketing.

THE CORN PIVOT (Sub-Indexes)

The composite tells you the national picture. Sub-indexes tell you which sectors. The key structural insight: corn is revenue for grain farmers but feed cost for dairy and livestock. When corn prices spike, grain farmers prosper while dairy farmers get crushed by feed costs. This creates a validated inverse correlation (r = -0.45)?When one goes up, the other goes down. An r of -0.45 means they move opposite about half the time. This is exactly what you'd expect: expensive corn is great if you're selling it, terrible if you're feeding it. between the grain and dairy sub-indexes.

Each sub-index computes a margin: sector revenue minus sector costs, both measured as z-scores?A way of saying "how unusual is today's price compared to history?" A z-score of +2 means the price is way above normal. A z-score of -2 means way below. Zero is average. This lets us compare apples to oranges — corn bushels to diesel gallons. against expanding historical windows. The grain sub-index uses corn/soy/wheat revenue against crude oil, diesel, fertilizer, farm machinery, and interest rate costs. Dairy uses milk/cheese/butter revenue against corn/soy feed costs plus energy and interest. Livestock uses cattle/hog revenue against the same cost structure.

Right now: Grain at 9.8 means row crop margins are at historic lows ($5 corn against elevated input costs). Livestock at 94.6 means cattle producers are in the best conditions in decades (record cattle prices). Same national composite — radically different realities depending on what you raise.

FORWARD OUTLOOK

The 4-quarter change in the Federal Funds rate?How much rates moved over the last year. If the Fed was at 5.3% a year ago and is at 4.5% now, that's a -0.8% change. Falling = good for farmers. Rising = trouble coming in 12-15 months. is the single strongest leading indicator?A signal that shows up BEFORE the problem does. Like seeing dark clouds before it rains. Rate hikes today show up as missed loan payments 4-5 quarters from now, because it takes time for higher payments to eat through farm cash reserves. of farm financial stress in our dataset. Rising rates predict higher delinquency 4-5 quarters later (r = -0.52). The current reading of 61.4 reflects easing from the 2023 rate peak — positive for farmers carrying debt over the next 12-15 months.

DATA SOURCES (16 FRED Series)

Corn (PMAIZMTUSDM) · Soybeans (PSOYBUSDM) · Wheat (PWHEAMTUSDM) · Crude Oil (POILWTIUSDM) · Diesel PPI (WPU057303) · Fertilizer PPI (WPU0652) · Farm Machinery PPI (WPU111) · Raw Milk PPI (WPU01610102) · Cheese PPI (PCU311513311513) · Butter PPI (WPU023201) · Cattle PPI (WPU0131) · Hog PPI (WPU013201) · Fed Funds (FEDFUNDS) · CPI (CPIAUCSL) · 10Y Treasury (GS10) · Ag Loan Delinquency (DRFAPGACBS)

All data from the Federal Reserve Bank of St. Louis (FRED). No proprietary data. No estimated inputs. History: Q1 2003 – present (92 quarters). Updated quarterly after FRED publishes complete quarter data.

Full methodology PDF includes complete sub-index formulas, weight tables, validation statistics, historical event validation, and honest limitations disclosure. Free API available for developers and media integrations.