From the Erie Canal to data centers: two centuries of rise, decline, and attempted reinvention of the United States’ industrial heartland, told through the numbers. Updated to 13 October 2026.
| Indicator | Before | After |
|---|---|---|
| Employed in U.S. manufacturing | 19.4 million (1979) | 12.7 million (2025) |
| Population of Detroit | 1.85 million (1950) | 649,000 (2025) |
| Share of the 8 core states in the U.S. population | 40% (1900) | 24% (2024) |
| U.S. share of global steel production | 46% (1950) | 4% (2024) |
What it is, and where the Rust Belt is
The name is a political label before it is a geographical one, and like all political labels its boundaries are fluid.
In September 1984, Democratic candidate Walter Mondale, speaking in Cleveland, accused Reagan’s trade policy of turning the industrial Midwest into a “rust bowl,” modeled on the Dust Bowl of the 1930s. Newspapers refined the phrase into “Rust Belt,” for symmetry with the “Sun Belt,” which Kevin Phillips had coined in 1969 for the fast-growing states of the South and Southwest. Before it rusted, the same region was called the Manufacturing Belt, the Factory Belt, or the Steel Belt: the factory belt that, between 1900 and 1960, produced most of the country’s steel, automobiles, rubber, and machinery.
The region has no official borders. The core is the quadrilateral between the Great Lakes and the Ohio River: western Pennsylvania, Ohio, southern Michigan, northern Indiana, the Chicago area, eastern Wisconsin, New York State west of Albany, and the northern tip of West Virginia. Around it lies a ring that shares its history: St. Louis and eastern Missouri, Baltimore, Kentucky along the Ohio, eastern Iowa; and, across the border, southwestern Ontario, from Windsor to Hamilton, which was Canada’s Detroit and Pittsburgh.
Geography explains why heavy industry was born here. Three ingredients lay within a few hundred kilometers of one another: iron ore from the Mesabi Range (Minnesota) and the Upper Peninsula of Michigan; coking coal from the Appalachians, between western Pennsylvania, West Virginia, and eastern Ohio; and water, which meant nearly free transport. The Great Lakes, the locks at Sault Ste. Marie (1855), the Erie Canal (1825) that connected them to New York and the Atlantic, and the Ohio, Monongahela, and Allegheny rivers. Ore came down by lake to the ports of Cleveland, Gary, and Buffalo; coal came up from the Appalachian valleys; and where the two flows intersected, blast furnaces rose. Straddling the late 19th and early 20th centuries, the region was also the center of the market: almost half of Americans lived within a day’s train ride.
Canals, iron, and immigrants
Everything begins with a ditch. The Erie Canal, opened on October 26, 1825, linked Lake Erie to the Hudson River: the cost of transporting a ton of goods between Buffalo and New York fell from about $100 to about $10. Buffalo became the continent’s grain port, Cleveland and Detroit grew as transshipment hubs, and the fertile agricultural belt of Ohio and Indiana finally gained a market. In 1855 the Sault Ste. Marie locks opened Lake Superior to navigation, and with it the iron deposits of Michigan and then Minnesota.
Pittsburgh had been working iron since the late eighteenth century, thanks to readily available coal. The breakthrough came in 1875, when Andrew Carnegie opened the Edgar Thomson Works in Braddock, the first major American steel mill built around the Bessemer converter: steel rails at a third of the cost, just as the railroads were laying tens of thousands of kilometers of track each year. In 1870 John D. Rockefeller founded Standard Oil in Cleveland; that same year B.F. Goodrich brought rubber processing to Akron. Chicago, already the continent’s leading rail hub, concentrated the slaughterhouses (the Union Stock Yards opened in 1865), McCormick reapers, and Pullman railcars. Milwaukee made machines and beer.
Labor arrived by ship. Poles, Slovaks, Hungarians, Italians, and Jews from Eastern Europe filled the working-class neighborhoods of Pittsburgh, Cleveland, and Chicago; western Pennsylvania spoke more languages than many European capitals. Industrial conflict was imported as well: in 1892 the Homestead strike, which ended in an armed clash with Pinkerton guards, shut unions out of the steel industry for forty years.
In 1901 J.P. Morgan merged Carnegie Steel and other producers into U.S. Steel, the first company in the world capitalized at more than one billion dollars; five years later U.S. Steel itself founded Gary, Indiana from scratch, a factory city on the southern shore of Lake Michigan.
The arsenal of democracy
The automobile transformed a region of steel mills into a region of assembly lines. Henry Ford founded his company in Detroit in 1903; the Model T came out in 1908, the same year William Durant put together General Motors in Flint. In 1913, in Highland Park, the moving assembly line cut the time needed to assemble a Model T from more than twelve hours to ninety-three minutes; in 1914 Ford raised pay to five dollars a day, twice the average, and invented the worker who buys what he produces. Around Detroit, a continental supply chain took shape: rubber in Akron (Goodyear and Firestone), glass in Toledo, components in Dayton, Kokomo, Saginaw, and Flint, flat-rolled steel in Gary, Cleveland, and the Mon Valley. From 1916 onward, the Great Migration brought about six million African Americans over fifty years from the rural South to Northern factories.
The Second World War did the rest. In December 1940 Roosevelt asked industry to become the “arsenal of democracy”; at Willow Run, near Detroit, Ford managed to produce one B-24 bomber per hour, and Chrysler built tanks in the suburb of Warren. After the war, the region reaped the dividends: the 1950 “Treaty of Detroit,” the five-year contract between GM and the UAW with pensions, health care, and cost-of-living adjustments, set the template for a working middle class that owned a home and a car. In 1950 Detroit had 1.85 million residents and one of the highest per-capita incomes among major U.S. cities; in 1953 nearly one American worker in three outside agriculture was employed in manufacturing.
The chart says two things. First, the curve is not a mountain with a single slope: U.S. manufacturing employment grew up to 1979 and then fluctuated around 17–18 million for two decades. The real collapse is concentrated in two windows, 2001–2003 and 2008–2009. Second, the Rust Belt began to rust before the national curve fell: between 1950 and 1979 factories did not disappear—they moved, to the suburbs, to the South, to California—and the factory cities of the Great Lakes were losing share even in the boom years.
Why it rusted
There is no single cause, and the causes reinforced one another. It is worth separating five, because political debate tends to pick just one.
1 · Foreign competition and oil
Japan and Germany rebuilt modern steelworks from scratch (basic oxygen furnace, continuous casting) while the U.S. integrated mills remained those of the 1920s. The 1959 steel strike, 116 days long, opened the door to imports. After the oil shocks of 1973 and 1979, Americans bought small, reliable cars: the share of imports in car sales rose from about 15% in 1970 to more than a quarter in 1980.
2 · Technology
Nucor’s electric mini-mills and those of the other new producers—almost all in the South—remelt scrap with a fraction of the workforce of integrated steelmakers. Robots and numerical control did the same in auto plants. Today, the United States produces about twice as many manufactured goods as in 1979 with one-third fewer workers.
3 · Labor costs and factory geography
The new plants went to the South’s “right-to-work” states, then to Mexican maquiladoras, and finally—after NAFTA (1994) and China’s entry into the WTO (2001)—overseas. The “China shock” estimated by Autor, Dorn, and Hanson amounts to between 2 and 2.4 million jobs lost between 1999 and 2011, concentrated in manufacturing counties that have never recovered.
4 · Macroeconomics
Volcker’s monetary tightening (1979–82) pushed the dollar to record levels and produced the 1981–82 recession: exports of steel and machinery became uncompetitive. Employment in steel fell from about 450,000 in 1979 to about 200,000 in 1990; today it is under 90,000.
5 · Cities
Suburbanization, white flight, and segregation hollowed out city centers; the 1967 riots (43 dead in Detroit) accelerated the exodus. The tax base shrank while pensions and debt remained: Cleveland defaulted on its debt in 1978, Detroit in 2013, and Flint ended up poisoning its own water in 2014.
The moment the region remembers as the beginning of the end has a date: Monday, September 19, 1977—the “Black Monday” of Youngstown—when Youngstown Sheet & Tube announced the closure of the Campbell Works and 5,000 layoffs in a single day; in the following five years the Mahoning Valley lost tens of thousands of jobs and half its reason for being. Cleveland’s default (1978) followed, as did the $1.5 billion in federal guarantees to save Chrysler (1979–80), the closure of U.S. Steel’s Homestead Works (1986), and the bankruptcies of LTV (2000) and Bethlehem Steel (2001). U.S. steel output, which had reached 137 million tons in 1973, was down to 68 in 1982.
The cities: eight parables
Factories were closing, but people were leaving even faster. The big core cities reached their population peak between 1930 and 1960 and since then have lost between half and two-thirds of their residents; the suburbs absorbed some, and the South and West the rest. Milwaukee and Chicago, with more diversified economies, held up better. Detroit, which in 2020 was at one-third of its 1950 level, returned to growth in 2023 for the first time since 1957.
Thirteen cities, from peak to the 2020 census. Residents within municipal boundaries.
| City | Peak year | Population at peak | 2020 | Change | Note |
|---|---|---|---|---|---|
| Detroit (MI) | 1950 | 1,849,568 | 639,111 | −65.4% | 649,095 residents in the 2025 estimate: third year of growth |
| Chicago (IL) | 1950 | 3,620,962 | 2,746,388 | −24.2% | Diversified economy, finance and services |
| Cleveland (OH) | 1950 | 914,808 | 372,624 | −59.3% | Cleveland Clinic is Ohio’s largest employer |
| St. Louis (MO) | 1950 | 856,796 | 301,578 | −64.8% | Western crown |
| Pittsburgh (PA) | 1950 | 676,806 | 302,971 | −55.2% | Universities and healthcare (“eds and meds”), robotics |
| Milwaukee (WI) | 1960 | 741,324 | 577,222 | −22.1% | Machinery, water and water technologies |
| Buffalo (NY) | 1950 | 580,132 | 278,349 | −52.0% | +6.5% between 2010 and 2020, thanks to immigrants and refugees |
| Toledo (OH) | 1970 | 383,818 | 270,871 | −29.4% | Jeep (Stellantis), glass |
| Akron (OH) | 1960 | 290,351 | 190,469 | −34.4% | Former world rubber capital, now polymers |
| Dayton (OH) | 1960 | 262,332 | 137,644 | −47.5% | Wright-Patterson Air Force Base |
| Flint (MI) | 1960 | 196,940 | 81,252 | −58.7% | GM’s hometown; water crisis since 2014 |
| Gary (IN) | 1960 | 178,320 | 69,093 | −61.3% | Gary Works, the largest integrated steel mill in the USA |
| Youngstown (OH) | 1930 | 170,002 | 60,068 | −64.7% | “Black Monday” 1977 |
The paradox: more output, fewer workers
Those who attribute everything to foreign competition must explain an uncomfortable fact: in real terms, U.S. manufacturing today produces about twice as much as in 1979. The following chart indexes output and employment to 1979. Part of the widening gap is automation; part is composition (computers and electronics, measured by quality, inflate the index: without them, output growth is much more modest); part is the end of low-productivity production. For the Rust Belt, the result is the same: even when factories come back, they come back with fewer people inside.
The core states
The decline has not been uniform. Indiana and Wisconsin have lost less and remain America’s most manufacturing-intensive states relative to employment; New York has lost almost half of its industrial workforce in twenty-five years, especially in Upstate. In all eight states, however, the trough was reached in 2009–2010, and since then there has been a partial rebound.
Manufacturing employment in the eight core states, 2000 → 2025. Thousands employed, annual averages, rounded values (BLS CES); share of total nonfarm employment; 2024 population in millions (Census).
| State | Portion in the core | Historic industries | 2000 | 2010 | 2025 | Change 2000–25 | Manufacturing share | Pop. 2024 |
|---|---|---|---|---|---|---|---|---|
| Ohio | Entire state | Steel, rubber, glass, autos | 1,020 | 620 | 660 | −35% | 12% | 11.9 |
| Michigan | Southern half | Automobiles, furniture | 895 | 470 | 590 | −34% | 13% | 10.1 |
| Pennsylvania | West and northeast | Steel, coal, aluminum, glass | 865 | 565 | 565 | −35% | 9% | 13.1 |
| Illinois | Chicago area and north | Steel, agricultural machinery, railroads | 870 | 570 | 550 | −37% | 9% | 12.7 |
| Indiana | North | Steel, auto components | 670 | 440 | 530 | −21% | 16% | 7.0 |
| Wisconsin | East | Machinery, beer, autos | 595 | 435 | 465 | −22% | 15% | 6.0 |
| New York | Upstate (west) | Steel, optics and photography, machinery | 750 | 455 | 415 | −45% | 4% | 19.9 |
| West Virginia | Northern panhandle | Steel, coal, chemicals | 76 | 50 | 46 | −39% | 6% | 1.8 |
| Eight states | 5,741 | 3,605 | 3,821 | −33% | 10% | 82.4 | ||
| United States | 17,265 | 11,528 | 12,700 | −26% | 8% | 340.1 |
Timeline, 1825–2026
The region that decides elections
Since 2016, the Rust Belt has moved back to the center of American politics—and politics has moved back to the center of its economy.
Pennsylvania, Michigan, and Wisconsin—the “blue wall” that Democrats took for granted—went to Trump in 2016, to Biden in 2020, and back to Trump in 2024, always by margins of one or two points. Ohio, once the quintessential bellwether state, has shifted steadily to the right. The result is that both parties have built their industrial policy with this region in mind: Biden’s CHIPS Act and Inflation Reduction Act (2022) steered semiconductor fabs and battery plants toward the Great Lakes; Trump opted for tariffs. On 3 November 2026, the midterm elections will once again be decided here: voters will elect the governors of Michigan, Ohio, Pennsylvania, and Wisconsin, as well as the senators from Michigan and Ohio.
Tariffs: who wins and who pays
Trade policy in Trump’s second term has had a turbulent year and a half, and the Rust Belt is its main target. The picture as of today is in the table.
Trade measures affecting the region. Status as of 13 October 2026.
| Measure | Legal basis | Timeline | Status | Effect on the region |
|---|---|---|---|---|
| Tariffs on steel and aluminum | Section 232 (national security) | 25% and 10% from March 2018 with exemptions; 25% without exemptions from March 2025; 50% from 4 June 2025; from 6 April 2026 applied to the full value of products (50% pure metals, 25% derivatives, 15% network equipment, 10% if made with American metal) | In force | Record margins for U.S. Steel, Cleveland-Cliffs, and Nucor; higher costs for cars, machinery, and household appliances, which employ far more people than steel mills |
| Tariffs on automobiles and parts | Section 232 | 25% from 3 April 2025 on imported cars, with discounts for USMCA content | In force | Pushes assembly back to Michigan, Ohio, Indiana, and Illinois (Stellantis, GM); makes parts from Canada and Mexico more expensive |
| “Reciprocal” and “fentanyl” tariffs | IEEPA (emergency powers) | Announced on 2 April 2025; struck down by the Supreme Court on 20 February 2026 (6–3, Learning Resources v. Trump): the law does not authorize tariffs. About 130 billion collected; refunds still to be defined | Struck down | Reduces uncertainty for Mexican and Canadian auto suppliers; opens disputes over refunds |
| Temporary global tariff | Section 122 | 10% on almost all imports, introduced in February 2026 immediately after the ruling; the law limits it to 150 days (until 24 July 2026) unless extended by Congress. In 2026 the administration also used Sections 301 and 338 | Temporary | A bridge after the invalidation of the IEEPA tariffs; does not apply to goods already hit by the 232 |
| USMCA review | U.S.–Mexico–Canada Agreement | Joint review launched in July 2026; 232 tariffs remain applicable even to compliant goods | Ongoing | Will determine the future of the integrated Detroit–Windsor–Toluca auto supply chain |
The paradox of steel tariffs is a paradox of numbers: in the United States about 140,000 people produce steel and aluminum, while more than six million work in industries that consume them—from cars to machinery to household appliances—many of which are located in the same states. U.S. Steel, now controlled by Nippon Steel, expects its best profit in 2026 since 2023; Ford, GM, and machinery makers, by contrast, have booked billions in costs. The April 2026 reform, which calculates the tariff on a product’s full value rather than only on its metal content, shifted the burden precisely onto processors.
Factories: chips, batteries, steel, cars
The season of announcements worth tens of billions, which began in 2022, has turned into a season of verification. The table summarizes where the main projects stand.
Major investments announced in the region, 2022–2026. Amounts in dollars as stated by the companies; status as of October 2026.
| Project | Location | Commitment | Status | Where we stand |
|---|---|---|---|---|
| Intel “Ohio One,” two semiconductor fabs | New Albany (OH), Columbus area | 28 billion (up from an initial 20); 3,000 jobs at full capacity | Delayed | Announced in January 2022 with a 2025 start-up; now the first fab is expected to be operational between 2030 and 2031, the second in 2032. Work at the site picked up again in early 2026 and Intel says AI chip demand exceeds supply, but it warns that if the 14A node fails to attract customers it will “probably” abandon the Ohio fabs |
| Micron, memory “megafab” | Clay (NY), Syracuse area | Up to 100 billion over twenty years | Ahead of schedule | Site work began in early 2026; in July 2026 vertical construction began more than a quarter ahead of plan; first production expected in 2030; 675 million already spent with New York State suppliers |
| Nippon Steel in U.S. Steel | Mon Valley (PA), Gary Works (IN) | $14.9 billion acquisition closed on June 18, 2025; 11 billion in promised investments by 2028, “golden share” for the U.S. government | Under way | Less than 200 million spent as of end-March 2026 and 3.2 billion in projects approved; in June 2026 the commitment for Mon Valley rose to 2–2.5 billion, with a new hot-rolling mill replacing an 87-year-old facility; a rebuild of Gary’s No. 14 blast furnace is planned |
| Stellantis, U.S. plan | Belvidere (IL), Toledo (OH), Warren and Detroit (MI), Kokomo (IN) | 13 billion over four years; more than 5,000 jobs | Under way | Announced on October 14, 2025, the largest in the group’s U.S. history: Belvidere reopening in 2027 for the Jeep Cherokee and Compass (3,300 jobs), a new midsize pickup in Toledo in 2028, large SUVs in Warren, a new four-cylinder engine in Kokomo starting in 2026 |
| Ford, LFP battery plant | Marshall (MI) | 3.5 billion initially, 1,700 jobs (down from 2,500) | Scaled back | Capacity cut from 400,000 to 230,000 vehicles per year; the end of federal incentives for electric cars (September 30, 2025) slowed the entire battery supply chain in the Great Lakes |
| Amazon Web Services, data centers | Salem Township and Falls Township (PA) | 20 billion, the largest private investment in Pennsylvania’s history | Under way | Announced in June 2025; one campus is connected directly to the Susquehanna nuclear power plant |
| Constellation, restart of Three Mile Island 1 (“Crane Clean Energy Center”) | Londonderry Township (PA) | 1.6 billion; twenty-year contract with Microsoft for 835 MW | Under way | Reactor shut down in 2019; restart planned for 2027, ahead of the original plan |
The new arrival: data centers
The industry growing fastest in the region doesn’t make objects. The Great Lakes region hosts about one-fifth of U.S. data centers: 525 in 2024 and another 224 planned by 2030, according to the Weldon Cooper Center at the University of Virginia. Ohio has more than a hundred operating, and the Columbus area—where Amazon, Google, and Meta already have campuses—is set to become the region’s second hub; projects planned in the state amount to more than 15 gigawatts of power, the equivalent of fifteen nuclear power plants. Pennsylvania, at Pittsburgh’s energy and artificial intelligence summit in July 2025, secured announcements totaling around 90 billion.
The reason is the same one that, a century ago, drew in steel mills: energy and water close at hand, industrial land already developed, power lines that used to serve factories that have since closed. The problem is the relationship between capital and labor. In Ohio, data centers generated about 22,300 temporary construction-site jobs in 2024 but only 4,500 permanent ones; the real spillover is property-tax revenue, which municipalities often forgo in order to attract them. In 2026 the first resistance emerged: in May, Cleveland rejected a $1.6 billion project; in Pennsylvania, a statewide moratorium is being debated; and in Ohio even Republican figures are calling for tax breaks to be removed.
Work and people, today
The latest data point to stabilization, not a boom. Nationally, manufacturing had lost 56,000 jobs between May 2025 and May 2026; from December 2025, the low point, it regained 72,000 through September 2026, almost all of them in durables (fabricated metals, machinery, transportation equipment), with some pull-forward effect from purchases due to tariffs. The October 2, 2026 employment report showed an economy that created only 29,000 jobs in a month, with manufacturing contributing 9,000 of them: proportionally, it is the sector holding up best.
On the demographic front, the two faces of the region are Detroit and Pittsburgh. Detroit grew for the third consecutive year, +5,060 residents in 2025 to 649,095, after persuading the Census Bureau to count refurbished homes and not subtract population for every demolition. The Pittsburgh metropolitan area shrank by 0.13% in 2025: it recorded 21,462 births and 29,258 deaths, the most negative natural balance among the forty largest U.S. urban areas, and international immigration—which in past years had offset the decline—fell to less than half amid the 2025–26 restrictions. It is the issue uniting the entire region: an aging population, few births, dependence on arrivals from abroad.
Three scenarios and a list of things to watch
The Rust Belt of 2035 will not be the Rust Belt of 1965, and it is not a given that it will remain the Rust Belt of 2010. The forces at play are clear enough to outline three scenarios.
Scenario A · Selective reindustrialization
Tariffs, what remains of the CHIPS Act, and AI-driven energy demand bring into the region a significant share of the capital the United States is investing in factories, power plants, and data centers. But these are capital-intensive industries: many billions, few jobs per billion, and good wages only for those with technical skills.
Winners: Columbus, Indianapolis, Grand Rapids, Pittsburgh, Madison, the Lehigh Valley. Manufacturing employment in the core climbs back above 4 million, but does not come close to the 5.7 million of 2000.
Scenario B · Two Rust Belts
The region is dividing in a stable way. Metropolitan areas with universities, major hospitals, and airports (“eds and meds”: UPMC, Cleveland Clinic, Michigan Medicine) keep growing and attracting young people; single-industry cities—Youngstown, Flint, Gary, Saginaw—keep losing residents and tax base, with public services cut to the bone.
This is the scenario closest to the trend of the last fifteen years. Politics matters: state incentives, county consortia, and local public-finance governance make the difference between Pittsburgh and Youngstown.
Scenario C · Rust returns
Stop-and-go tariffs drive up costs for metal fabricators, demand for cars falls, the electric transition stalls halfway—leaving half-empty battery plants—the data-center bubble deflates before the power plants are built, and yet another change of administration once again overturns incentives.
With immigration at a minimum, demography does the rest: a region that loses residents also loses the base for any kind of revival. It is the scenario local labor markets have already lived through in 1982 and in 2009.
Tailwinds
- Fresh water: the Great Lakes hold about 21% of the planet’s surface fresh water, protected by the Great Lakes Compact (2008); a structural advantage over Arizona and Texas for industry, agriculture, and data-center cooling.
- Climate: cooler summers and no hurricanes; Buffalo, Duluth, and Cleveland market themselves as “climate havens.”
- Low-cost housing and land, existing electric and rail infrastructure, ports on the lakes.
- Widely distributed mechanical skills, top-tier engineering schools (Carnegie Mellon, Michigan, Purdue, Illinois, Ohio State) and a robotics ecosystem in Pittsburgh.
- Politics: as long as elections are decided here, both parties have an interest in investing in it.
Headwinds
- Demography: more deaths than births in almost all metropolitan areas; dependence on immigration just as it is being restricted.
- Regulatory uncertainty: three tariff regimes in two years, electric-vehicle incentives introduced in 2022 and repealed in 2025, the CHIPS Act renegotiated.
- Input costs: with 50% tariffs, American steel costs more than European and Asian steel; exporters of machinery and vehicles lose competitiveness.
- Power grid: the gigawatts promised to data centers do not exist yet; PJM, the area’s grid, already has record-high capacity prices.
- Fiscal legacy: municipal pensions, water systems, and schools sized for cities twice as large.
What to watch over the next twelve months
- 3 Nov 2026 — Midterm elections: governors in Michigan, Ohio, Pennsylvania, and Wisconsin; senators in Michigan and Ohio; the outcome will shape tariffs and incentives for the following two-year period.
- 2026–27 — Outcome of the USMCA review: if Section 232 tariffs remain on compliant goods, the Detroit–Ontario–Mexico auto supply chain will keep reorganizing; if they are lifted, Belvidere and Toledo will go back to competing with Toluca.
- late 2026 — Intel’s decision on the 14A node and, by extension, its Ohio fabs; potential outside customers for the foundry.
- 2027 — Reopening of Belvidere (Stellantis), restart of the Crane reactor (formerly Three Mile Island), progress on Micron’s construction sites in Clay.
- 2027–28 — Pace of the $11 billion pledged by Nippon Steel: the new Mon Valley rolling mill is the credibility test for the deal.
- May 2027 — Census Bureau 2026 demographic estimates: if Detroit grows for a fourth year and Pittsburgh stops shrinking, Scenario B loses force.
- ongoing — Monthly BLS reports on manufacturing employment: the psychological threshold is 12.8 million, the pre-pandemic level; below 12.5, Scenario C becomes likely.
Why it matters
The Rust Belt is the laboratory in which the United States is deciding whether reindustrialization by decree (tariffs, incentives, data centers) can bring jobs back to places that have been losing them for fifty years. The numbers show that output never disappeared: the workers did—and today demography weighs more than factories. As long as the “blue wall” of Pennsylvania, Michigan, and Wisconsin decides elections, both parties will keep writing American industrial policy with these eight states in mind; and whatever works or fails here, from steel to chips, sets the template for the rest of the industrial West.
Where the numbers come from
The historical series are rounded annual averages; state-level data through 2025 are estimates based on preliminary BLS data (rounded to the nearest 5,000); the production index is recalculated on a 1979 base from rounded values. The most recent figures come from the cited sources, consulted on 13 Oct 2026.
- Bureau of Labor Statistics, The Employment Situation, September 2026 (2 October 2026) and the Current Employment Statistics series (manufacturing, NAICS).
- U.S. Census Bureau, decennial censuses 1900–2020 and Vintage 2024/2025 estimates; City of Detroit, press release on the 2025 estimates; Bridge Michigan.
- University of Pittsburgh UCSUR, Census Estimates Brief, March 2026 (Pittsburgh metropolitan area).
- USGS, Mineral Commodity Summaries (iron and steel); AISI, weekly raw steel production, cumulative 2025 figure.
- Federal Reserve, Industrial Production and Capacity Utilization (G.17), manufacturing index.
- Congressional Research Service, Section 232 Tariffs on Steel and Aluminum (IN12519, updated as of 17 September 2026); White & Case, changes to the Section 232 tariffs of 2 April 2026.
- Jones Day, United States Supreme Court Invalidates IEEPA Tariffs (February 2026); Thomson Reuters, note on the ruling Learning Resources v. Trump.
- Axios, U.S. Steel, one year after the sale to Nippon Steel (30 June 2026); Jiji Press/Adnkronos, investment in the Mon Valley (9 June 2026); Reuters on the $11 billion plan (August 2025).
- Stellantis, 14 October 2025 press release, via Supply Chain Dive; AP/CBS Detroit on the Marshall battery plant.
- Data Center Dynamics, Intel further delays Ohio fabs; NBC4/AOL on Intel Ohio One (January 2026); W.Media, Micron begins vertical construction at Clay (July 2026).
- WYSU/Ohio Newsroom, Weldon Cooper report on Great Lakes data centers (January 2026); NBC Philadelphia, Amazon, $20 billion in Pennsylvania; Baxtel, coverage of projects in Ohio and Pennsylvania (2026).
- D. Autor, D. Dorn, G. Hanson, The China Shock, Annual Review of Economics, 2016; Encyclopaedia Britannica, “Rust Belt”; CEPR, September 2026 jobs report preview.