Midterms 2026
Democrats Won Congress: What Happened to Manufacturing After Nine Opposition Midterm Wins?
Nine flips, two parties and eight decades of factory data. The record cannot say which party is better for factories, and the reasons are worth knowing before anyone tells you otherwise.
In the year after Republicans took Congress in 1994, American manufacturing output rose 4.2%. In the year after Democrats took it in 2006, output rose 3.7%, and a year after that it was falling. In the year after Democrats took the House in 2018, output fell 3.5%.
Three flips, two parties, three different outcomes. On Tuesday the party out of the White House won both chambers for only the fifth time since World War II, CBS News noted, and this week boards and lenders will ask what it means for the plan. The government's factory data, some of it reaching back to 1939, gives a careful answer: comparisons like these cannot isolate the effect of an election. What they can show is what was happening in factories before and after each one.
What we measured
Between 1946 and 2022 the party out of the White House took at least one chamber at a midterm nine times; this story calls those nine the flips. Each comparison runs from October of the election year, the last full month before the vote, to 12 and 24 months later. Each is set against every other window of the same length in the archive, and against the other midterms in the archive, the ones in which the opposition took no chamber. One of those, 2002, was not a year in which nothing changed: the president's own party took the Senate. Not every series reaches back to every flip.
| Measure, with its series | Archive starts | Flips covered |
|---|---|---|
| Factory workweek (BLS via FRED, AWHMAN) | 1939 | 9 of 9 |
| Manufacturing output (Federal Reserve, IPMAN) | 1972 | 7 of 9 |
| Factory payrolls and hourly earnings (BLS CES) | 1990 | 6 of 9 |
| Producer prices (BLS PPI) | 1990 | 6 of 9 |
The nine flips, one row each
Change over the 12 months from October of the election year. Output is the Federal Reserve's manufacturing industrial production index; the workweek is average weekly hours of production and nonsupervisory workers in manufacturing. All three series are seasonally adjusted.
| Year | Chamber taken | Taken by | President's party | Output | Payrolls | Workweek |
|---|---|---|---|---|---|---|
| 1946 | Both | Republicans | Democratic | n/a | n/a | +0.2% |
| 1954 | Both | Democrats | Republican | n/a | n/a | +3.3% |
| 1986 | Senate | Democrats | Republican | +8.1% | n/a | +1.2% |
| 1994 | Both | Republicans | Democratic | +4.2% | +0.4% | -1.4% |
| 2006 | Both | Democrats | Republican | +3.7% | -2.2% | +0.2% |
| 2010 | House | Republicans | Democratic | +3.2% | +1.9% | +0.5% |
| 2014 | Senate | Republicans | Democratic | -0.9% | +0.8% | -0.7% |
| 2018 | House | Democrats | Republican | -3.5% | 0.0% | -1.7% |
| 2022 | House | Republicans | Democratic | -1.5% | -0.4% | -0.5% |
Output
Manufacturing output was higher a year later after four of the seven flips the archive covers. The median gain, 3.2%, sits a little above the 2.2% median for every 12-month stretch since 1972 and above the 1.3% median after the other midterms. Two years out the order reverses: the median after a flip is a decline of 1.4%, against gains of 3.4% for the archive as a whole and 4.5% after the other midterms. A result that changes sign with the window is a reason for caution, not a finding about elections.
Jobs: a pattern for each party
Jobs are where the data hands something to each side, and both readings deserve to be printed before they are taken apart. Factory payrolls rose in the year after three of the four Republican-won flips the archive covers and slipped after the fourth, 2022. They did not rise after either Democratic-won flip. A Republican could print that column. A Democrat could answer with 1986, when output rose 8.1% in the year after Democrats took the Senate, the strongest year in the table.
Neither reading holds up. There are two cases on one side and four on the other. Both Democratic-won windows opened with a downturn already under way: housing starts had fallen 27.8% in the year before the 2006 election, and the trade war that weighed on factories through 2019 began before the 2018 vote. The table also reads upside down. Every Republican-won flip came under a Democratic president, and every Democratic-won flip under a Republican one, so the same numbers can be made to say that factories do better under Democratic presidents.
Wages, in nominal dollars
Average hourly earnings of production and nonsupervisory workers in manufacturing rose in the year after all nine midterms since 1990, with median gains of 2.7% after a flip and 3.0% after the other midterms. These are nominal wages, before inflation, so they say little about purchasing power. What they show is a series that kept its trend whatever the result.
The other midterms
The sharpest contrast in the data is the least useful. In the other three midterms since 1990, factory payrolls fell every time: in 1990 and 1998, when control of Congress held, and in 2002, when the president's own party took the Senate. Those years include two recessions and the long decline in factory employment after 2000, and three cases cannot separate politics from the business cycle.
2018, the closest parallel
The nearest parallel to this week is 2018: the same president, tariffs in force, a Democratic House. Over the year that followed, as output slid, new orders fell 5.9%, the factory hiring rate fell 22.6% and the steel mill price index fell 12.8%. An escalating trade war and a global slowdown were under way at the same time, which is why one year cannot be credited to one election.
Where this cycle starts
For a plant, the useful comparison is not the party but the conditions in place when the gavel changed hands. In 2018, orders were rolling over. This fall the demand signals are middling. The workweek, at 42.0 hours in September 2026, is the longest since January 2019, and the Empire State manufacturing survey read 25.5 in October 2026, its strongest since December 2021. But capacity utilization in manufacturing, at 75.9% in September 2026, sits at its average since 2010 and below its 1972 to 2026 average of 78.1%.
The cost signals are not middling. Of the 13 producer price indexes in our archive, an overlapping set rather than a designed basket, 11 were higher in September 2026 than a year earlier, with a median rise of 8.7%, and diesel is $6.29 a gallon.
The cost signals are not middling.
In October 2027, this year's flip will get a 10th row in the table. The row will not say whether the election moved anything. Read alongside the producer price indexes, it will show whether the costs rising now were still rising a year later.
Manufacturing industrial production, every monthly observation since 1972, with its latest reading and the source release. Follow the output series
Method: How these figures were made
Nine flips: 1946, 1954, 1986, 1994, 2006, 2010, 2014, 2018 and 2022, the midterms between 1946 and 2022 in which the party out of the White House took at least one chamber (five by Republicans, four by Democrats). Windows run from October of the election year to 12 and 24 months later; medians compare every same-length window in each archive and the midterms with no opposition gain (1990, 1998 and 2002). Series: Federal Reserve G.17 IPMAN and MCUMFN; BLS AWHMAN, CES3000000001 and CES3000000008; BLS commodity PPIs; the New York Fed's Empire State survey; all via FRED. The event-study tables are fixed; the current readings, the PPI count and the diesel price are live. The capacity utilization averages (78.1% for 1972 to 2026, 75.9% since 2010) are fixed anchors. Fixed anchors were computed from the MFG Calcs archive as committed on Nov. 4, 2026 (commit e93297fa) and are stated with their windows; the site regenerates its data on its regular refresh schedule, so later readings may differ.
Further reading
- Industrial Production and Capacity Utilization, G.17, Board of Governors of the Federal Reserve System.
- Manufacturing industrial production, IPMAN, average weekly hours, AWHMAN and Empire State survey, GACDISA066MSFRBNY, FRED, Federal Reserve Bank of St. Louis.
- Election Day 2026 live updates, CBS News, Nov. 3 and 4, 2026.
- Manufacturing industrial production, average weekly hours and capacity utilization, the live series behind this story.
Published 2026-11-04.