On April 19, 1935, Hugh Hammond Bennett was testifying before the Senate Public Lands Committee, and he was stalling. Bennett ran the Soil Erosion Service. He'd spent years arguing that America was losing its topsoil at catastrophic rates. He also knew something the senators didn't: a dust storm that had lifted off the Southern Plains two days earlier was rolling east. So he repeated points. He circled back to data he'd already covered. He kept talking until the sky outside the Capitol windows went dark.
"This, gentlemen, is what I'm talking about."
Congress passed the Soil Conservation Act within weeks. Bennett became the first chief of the new Soil Conservation Service. It's a perfect story. The right man, the right moment, nature itself arriving on cue. It gets retold in every account of the Dust Bowl for a reason: it feels like the turning point.
The turning point that shaped what followed was quieter. It happened two years later, in a room where nobody was looking out the window.
The Design Choice
In February 1937, Roosevelt sent every state governor a draft of the Standard State Soil Conservation Districts Law. Not federal legislation. A template for states to adopt, creating local conservation districts as subdivisions of state government. Arkansas enacted it first, on March 3. By 1948, all forty-eight states had followed.
The law was drafted by a young USDA attorney named Philip Glick. In interviews decades later, Glick explained the calculus. The original vision included mandatory land-use regulations. But coercion was politically impossible. States'-rights resistance was fierce. And Glick identified a more practical problem: without sufficient funding to help landowners comply, mandatory regulations would punish farmers who wanted to conserve but couldn't afford to, right alongside those who simply refused.
So Bennett and Glick built the program around voluntary compliance and local control. Districts would be governed by boards of supervisors, elected by landowners. The boards would coordinate technical assistance, distribute federal cost-sharing, decide which farms received help first. Bennett called it:
"One of the best, and certainly the most promising, of the devices yet invented by man for dealing democratically and effectively" with the crisis of eroding land.
He wasn't being cynical. He faced a real constraint and found a workable solution. The soil needed saving now. Mandatory regulations would have died in state legislatures. Voluntary compliance with local governance was arguably the only version that could pass. And it worked, ecologically. Soil loss slowed.
What I keep coming back to is what "democratically" looked like once the boards were seated.
Who Governed
The boards were elected by landowners. In practice, the largest operators shaped the districts. A board dominated by a county's biggest landholders set the district's work plan, decided whose terraces got engineered first, whose applications for federal cost-sharing moved to the top of the pile. Technical assistance and conservation dollars flowed through these bodies, and the bodies were accountable to the people with the most acreage.
Topsoil stabilized across the Plains. What happened to the people on that soil moved in one direction. Farm numbers in the Great Plains peaked at nearly 1.7 million in 1935. By 1992, 646,000 remained. In Oklahoma and Texas, total farm numbers dropped 23 percent between 1935 and 1945, but tenant farmers were cut nearly in half. In counties across southwestern Kansas, places like Haskell County where the dust storms had been worst, average farm size more than doubled by 1950.
Consolidation had many drivers. Mechanization, drought, credit structures, the basic reality that homestead-sized parcels were never ecologically viable on the Plains. The district system was one among them. Historian Pete Daniel has documented how locally elected agricultural committees across USDA programs used what he calls "democratic elements" to obscure systematic exclusion. Ballots sent only to owners. Information withheld from tenants and sharecroppers. Between 1940 and 1974, the number of African American farmers fell from 681,790 to 45,594. The governance architecture Bennett pioneered, locally elected boards weighted toward landowners and replicated across USDA programs, wasn't the only instrument of that collapse. But it was one of the rooms where it happened.
The Rhyme
Today, when a flood destroys a neighborhood, the federal government can buy damaged properties through FEMA's Hazard Mitigation Grant Program. Over 43,000 properties purchased since 1989. Homeowners cannot apply directly. Local governments apply on their behalf, deciding which properties to prioritize, which neighborhoods to target, which communities justify the 25 percent local cost match.
A 2020 study found what you might already suspect: whiter counties were more likely to gain access to buyout assistance, while homeowners in neighborhoods of color were more likely to accept offers, "making nonwhite neighborhoods in otherwise white counties the areas of greatest demolition."
Both programs were built around real political constraints, real problems of scale and legitimacy. Both rely on local institutions that reflect the communities they govern. The compromise that made each program viable is the design choice that shaped who it served.
Bennett needed local buy-in to save the soil. He got it by giving local boards control. Ninety years later, FEMA needs local participation to move people out of floodplains. It gets it the same way. The dust cloud made the case for action. The board meeting, the one nobody remembers, decided who the action was for.
- Bennett's structure under threat: The White House has proposed eliminating the NRCS's entire Conservation Technical Assistance division, the direct descendant of Bennett's program, even as 2025 dust storms killed twelve people in Kansas and Texas.
- The cost-share equity gap: A 2026 game-theoretic analysis found that FEMA's current 75/25 buyout cost-sharing achieves strikingly low equity scores, and that an equity-weighted cost-sharing model could match outcomes at 25 percent lower cost by directing resources to lower-income communities.
- Soil loss hasn't stopped: Despite decades of conservation districts, U.S. croplands still lose at least twice as much soil annually to erosion as the Great Plains lost during the peak of the Dust Bowl, according to the Union of Concerned Scientists.
- The man who drafted it: Philip Glick's extended oral history interviews with NRCS historian Douglas Helms remain the most detailed account of why the 1937 law was designed the way it was, including his candid explanation of why mandatory compliance was shelved.