Farm consolidation over 40 years was steady and structural; land values rose far more but moved in volatile boom-bust cycles largely independent of the consolidation trend

Summary

Since 1982 the number of U.S. farms fell from 2.24 million to 1.90 million (-15.2%) through 2022, a slow, steady decline with no sharp inflection. But the average farm size barely moved (440 to 463 acres, +5.2%) because total farmland also shrank (-10.8%, 987M to 880M acres) -- the mean masks the real consolidation story, which is a redistribution: since the 1980s the number of very large crop farms (2,000+ cropland acres) has nearly doubled while midsize farms (100-999 cropland acres) fell by roughly half, and the midpoint acreage of both cropland and harvested cropland farms roughly doubled between 1987 and 2012 (about 650 and 488 acres, respectively, to about 1,200 acres each). Sales are far more concentrated than acreage: by 1997 just 2% of farms accounted for half of all farm sales (versus 17% of farms in 1900). Land values, meanwhile, rose far more than any consolidation metric -- nominal farm real-estate value per acre rose from $823 (1982) to $4,080 (2023), +396%, and even after stripping out inflation rose 86% in real terms. But the land-value series is NOT a smooth complement to the consolidation trend: it crashed roughly 41% in real terms during the 1980s farm-debt crisis (1982-87) while the farm count kept declining anyway, then boomed in 2004-2014 and again in 2021-2023. Across the 8 census years, farm count and land value per acre show only a moderate, statistically insignificant negative correlation (r=-0.50, p=0.20, n=8), and mean farm size shows essentially none (r=-0.04). Consolidation looks like a steady structural trend (mechanization, economies of scale, exit of small operations); land values look like a cyclical asset-price series driven by interest rates, commodity prices and non-farm demand for land -- the two are directionally consistent over four decades but not tightly coupled year to year.

1. How much has ownership consolidated?

(Most recent USDA survey, 2024, not a census year: ~1.88M farms, 876M acres, 466-acre average -- consistent continuation of the trend.)

Because both the number of farms AND total farmland acreage shrank at similar rates, the population-wide average farm size is nearly flat over 40 years -- a genuinely counter-intuitive result given how often 'farms are getting bigger' is stated as a simple fact. The actual consolidation is a shift within the size distribution, documented by USDA ERS: (1) the number of very large crop farms (2,000+ cropland acres) has nearly doubled since the 1980s while midsize farms (100-999 cropland acres) fell by nearly half; (2) the cropland midpoint acreage (the farm size at which half of all cropland is on larger farms and half on smaller ones) roughly doubled from 1987 to 2012, from about 650 acres to about 1,200 acres for total cropland and from about 488 to about 1,200 acres for harvested cropland; (3) production/sales concentration is far starker than acreage concentration -- by 1997, just 2% of U.S. farms accounted for half of all farm sales and 15% of farmland, versus 17% of farms and 43% of land in 1900.

2. Has that tracked with land values?

Farm real estate value per acre (USDA ERS/NASS annual series, national average, land+buildings) rose from $823 in 1982 to $4,080 in 2023 nominal (+396%), or from $2,193 to $4,080 in real 2023 dollars (+86%, GDP deflator). This is a much larger multiple than any consolidation measure moved over the same span. But the land-value path was highly cyclical, not a steady complement to the smooth consolidation trend: real land values fell about 41% from 1982 to a 1987 trough (the 1980s farm-debt crisis) even as the number of farms kept falling through that same period; values then rose steadily from the late 1980s through 2008, dipped slightly in the 2008-09 financial crisis, resumed a strong climb through the 2012-14 commodity boom, largely plateaued 2015-2020, then jumped again in 2021-2023 (higher commodity prices, low interest rates turning to inflation, non-farm/investor and renewable-energy demand for land).

Testing the relationship directly at the 8 census-year grain: correlation between number of farms and $/acre land value is r=-0.50 (p=0.20, not significant at n=8) -- consistent with the plausible mechanism that scarcer, pricier land pushes remaining farmers toward larger, fewer operations, but not a statistically robust finding given how few independent data points a 40-year census series provides and given that both series share a common long-run time trend (a classic confound for any 2-series correlation over time). Correlation between mean farm size and $/acre land value is r=-0.04 -- essentially no relationship, because mean farm size is the least informative of the three consolidation measures (see Section 1).

Bottom line: land values and farm consolidation moved in the same broad direction (fewer/bigger operations, much pricier land) over 40 years, but they are not tightly coupled -- consolidation looks driven mainly by structural forces (mechanization, economies of scale, small-farm exit) that continued through a land-value crash, while land values look driven by cyclical macro and commodity forces that produced booms and a bust the consolidation trend never mirrored.

Every query behind this report

1 warehouse call ran in this session, in order. Each is reproducible against the same snapshot.

SELECT DISTINCT category, subcategory, unit_desc FROM ag.ers_farm_income WHERE state='US' AND (unit_desc ILIKE '%per farm%' OR unit_desc ILIKE '%acre%' OR category ILIKE '%asset%' OR category ILIKE '%value%' OR subcategory ILIKE '%land%') ORDER BY 1,2