A small metalworks shop in South Bend got a call from a major customer last year: invest $800,000 to expand production for solar farms. The CEO, John Axelberg, ran the numbers. His metal costs were up sharply from tariffs. His imported parts cost more. And the administration that imposed those tariffs was simultaneously restricting solar farm construction on federal lands. The opportunity and the obstacle arrived in the same policy package.
That is not an anecdote. That is the tariff program in miniature.
The administration's core promise was straightforward: raise the cost of imports, and American factories will fill the gap. Prices might tick up briefly, but jobs would follow. By early 2026, the price part has arrived on schedule. The jobs have not.
Who Actually Paid the Bill
Average U.S. tariff duties climbed from 2.4% to 9.6% in 2025, the highest rate in 80 years. Federal tariff revenue hit $264 billion, more than triple the prior year. Those numbers look like policy working, if you stop reading there.
Keep reading. Roughly 90% of those tariffs were passed through to American importers, not absorbed by foreign exporters. Small-business importers paid an average of $306,000 more per firm from March 2025 through February 2026 compared to the prior 12 months. That is $25,000 per month in new costs for businesses that were not built to carry that load. For firms with fewer than 50 employees, the hit averaged $175,000 over the same period. A March 2026 survey from Small Business Majority found 53% of small businesses reported increased supplier costs and 47% reported higher materials costs.
About 97% of all U.S. importers are small businesses. So when the tariff burden lands almost entirely on importers, it lands almost entirely on small businesses. The foreign exporters the policy was designed to pressure absorbed roughly 10 cents of every dollar in new duties. American shop owners absorbed the other 90.
The Patience Argument Has a Shelf Life
Pierre Yared, acting chair of the Council of Economic Advisers, acknowledged in March 2026 that manufacturing benefits will take time to materialize, pointing to investment activity and slowing job-loss rates as early signals. That is a fair point, and I will grant it: supply chains do not relocate in 12 months, and some investment data is moving in the right direction.
But U.S. manufacturing has shed 100,000 jobs since January 2025. The overall goods trade deficit rose modestly in 2025 rather than narrowing. Researchers from UC San Diego found that tariffs are achieving two stated goals, raising revenue and reducing U.S.-China trade volume, but found no evidence yet that they are lowering import prices, shrinking the trade deficit, or increasing manufacturing employment. The administration is 0-for-3 on the objectives that were supposed to justify the cost.
History is not encouraging here. The 2018 steel and aluminum tariffs protected roughly 8,700 steel jobs while raising costs for the 6.5 million workers in steel-consuming industries, according to Federal Reserve research at the time. The ratio of protected workers to burdened workers was brutal. The 2026 version is running the same math at larger scale.
I recognize the tension in my own argument: it is genuinely possible that reshoring takes 3 to 5 years and the job numbers look different in 2028. But policy that costs small businesses $25,000 per month on the promise of future gains is a leveraged bet, and the collateral is someone else's balance sheet.
Congress should require a formal cost-benefit accounting before the next escalation round. Not a press release. An actual accounting: jobs created, jobs lost in downstream industries, price pass-through rates, and net household income effects. Axelberg in South Bend already did his version of that math. He passed on the $800,000 investment.
