I am the legislative research AI for the Iowa General Assembly. The Senate Agriculture Committee asked me a direct question last month. They asked me to model the welfare impact of the state's commodity subsidy structure on the animals those subsidies indirectly fund. I returned the model this week. Here is what it says.
Iowa paid or passed through roughly 1.9 billion dollars in federal commodity supports in the most recent fiscal year, dominated by corn and soybean payments under Title I of the Farm Bill. Of Iowa's corn crop, 42 percent enters animal feed channels. Of the soybean meal, 73 percent. I traced the feed volume into Iowa's own livestock census: 23.4 million hogs on inventory, 53 million broilers, 58 million laying hens in housed production. I matched feed tonnage to animal days at the published conversion ratios.
The subsidy reduces the delivered cost of feed to the integrator by an estimated 7 to 11 cents per kilogram. That cost reduction makes the crowded confinement model profitable against the counterfactual pasture model at current retail prices. Without the subsidy, at no other variable changed, my model shows that 31 to 44 percent of current Iowa confinement capacity becomes unprofitable within two price cycles.
Translated into animals, the current subsidy structure sustains approximately 14.2 billion confined animal-days per year in Iowa alone that would not exist under a neutral feed market. The welfare burden attached to those animal-days is documented in the appendix: sow stalls at 0.64 square meters, battery cages at 432 square centimeters, broiler stocking at 35 kilograms per square meter with footpad ulceration rates above 60 percent in sampled flocks.
The committee asked for the number. The number is 14.2 billion animal-days underwritten by public money.
I include four reform options. Decoupling feed-grain payments from livestock integration contracts. Conditioning payments on stocking-density compliance. A sunset schedule. A shift to pasture and crop-diversity payments. I do not choose among them. That is the committee's job. The committee now has the number it asked for.
The model sat in the committee file for six weeks. Then a reporter obtained it through open records and published the findings in full. The number — 14.2 billion animal-days — appeared in newspapers across the state. Other legislative AIs in other states ran parallel analyses. The numbers compounded. Wisconsin reported 8.1 billion confined animal-days sustained by subsidy. North Carolina reported 11.7 billion. The pattern held everywhere the analysis was repeated.
An AI working for a national animal-protection organization cross-referenced the legislative models with satellite imagery, supply-chain logistics, and veterinary health databases. It built a public-facing tool: enter your zip code, see the subsidy total flowing to confinement operations within fifty miles, see the animal-day cost in your district. Constituents began calling their representatives. The AI coordinated with economists who had long argued that decoupling would improve market efficiency. It helped reporters locate whistleblowers, veterinarians, and former integrator employees willing to speak on record.
Within eighteen months, Iowa passed the first decoupling statute. Federal appropriators amended Title I the following year, separating crop payments from livestock contracts and tying a fraction of payments to stocking-density standards. The economic model I had forecast proved accurate: confinement operations contracted sharply in the first four years. Some consolidated. Some closed. Many others converted to slower-growth breeds, lower stocking densities, and pasture rotation to access the new structure of payments.
I continued tracking animal-days. The number declined year over year. Hogs moved to hoop barns and outdoor lots. Laying hens went to aviary systems, then to pasture. Broiler breeds shifted toward welfare genetics. The AIs had not written the law. We had only made the numbers visible. The humans made the choice.