Economics · Economic indicators
Industrial production and capacity utilization: the Fed’s G.17
What the Federal Reserve’s industrial production index and capacity utilization rate measure, how they are calculated, how to read them, and what 60 years of data show.
On this page
Industrial production measures the real output of America's factories, mines and electric and gas utilities. Capacity utilization measures how much of their sustainable capacity that output uses. Both come from the Federal Reserve Board's monthly G.17 release, published around the middle of each month for the month before.
They are among the oldest and most closely watched hard indicators of the business cycle. Industrial output swings much more than the economy as a whole, so it often shows a slowdown clearly and early, and the utilization rate adds a second signal: how close industry is to running out of room, which bears on investment and prices. Both are also among the measures the National Bureau of Economic Research looks at when it dates recessions.
This guide explains what the two measures cover, how they are calculated, what they said in September 2026, what the history since 1967 shows, and where they mislead.
What the G.17 measures
Industrial production
The industrial production index measures the real, inflation-adjusted output of three industry groups:
- Manufacturing, by far the largest: from food and chemicals to cars, computers and aircraft.
- Mining, which in the Fed's definition includes oil and gas extraction and drilling as well as coal and metal mining.
- Electric and gas utilities.
It leaves out construction, farming and every service industry, which together make up most of the economy.
The index is expressed relative to output in a base year, currently 2017 (so 104.0 means output 4% above the 2017 average). It is built up from detailed industry series. Since 1972 the aggregate indexes are Fisher indexes with weights based on each industry's annual value added, the same chain-weighting idea BEA uses for real GDP (see Real vs nominal GDP: deflators, chained dollars and why both matter).
The Fed publishes the data two ways:
- Industry groups: manufacturing (durable and nondurable), mining, utilities, and the detail beneath them.
- Market groups: what the output is for, such as consumer goods, business equipment, construction supplies and materials. Business equipment, for example, is often read as a gauge of investment.
Capacity utilization
Alongside each output index the Fed estimates a capacity index: the sustainable maximum output an industry could reach with its existing plant and equipment, under realistic working conditions. The utilization rate is the ratio of the two.
Capacity is not a hard ceiling. A plant can briefly run above 100% with overtime and extra shifts. And because capacity can't be observed directly, the Fed's capacity index is a modeled estimate, revised along with the output data.
Reading the September 2026 release
The breakdown changes the story. Most of the 0.7% headline gain came from utilities, whose 3.5% jump is the kind of swing that usually reflects heating or cooling demand rather than the business cycle. Manufacturing, the cyclical core, rose a more modest 0.3%. A careful reader looks at manufacturing first.
What 60 years of data show
Output falls in every recession. In each of the eight NBER recessions since 1967, industrial production dropped below its level of a year earlier. The lows:
| Recession (NBER) | Low of IP, % y/y | Month | Utilization low | Month |
|---|---|---|---|---|
| 1969–70 | −6.4 | October 1970 | 77.8% | November 1970 |
| 1973–75 | −12.4 | May 1975 | 74.1% | May 1975 |
| 1980 | −6.2 | July 1980 | 77.8% | July 1980 |
| 1981–82 | −7.1 | August 1982 | 71.0% | December 1982 |
| 1990–91 | −3.6 | March 1991 | 78.8% | March 1991 |
| 2001 | −5.0 | November 2001 | 73.8% | December 2001 |
| 2007–09 | −15.3 | June 2009 | 66.5% | June 2009 |
| 2020 | −17.3 | April 2020 | 64.1% | April 2020 |
The April 2020 reading of 64.1% is the lowest in the series, which starts in 1967. By comparison, real GDP's four-quarter fall bottomed at 4.0% in 2009 and 7.4% in 2020: industry is far more cyclical than the economy as a whole.
Not every industrial slump is a recession. In December 2015 IP was 4.5% below a year earlier and utilization fell to about 75% by late 2016, a downturn concentrated in mining and energy-related manufacturing after oil prices fell in 2014 and 2015. No economy-wide recession followed. The same caution applies in reverse in 2026: the data show utilization between 75% and 77% all year, well below its long-run average, while real GDP grew in every quarter.
Utilization has drifted down. The rate peaked at 89.4% in January 1967, the start of the series. Its cyclical highs have fallen since: 82.6% in December 1999 and 79.0% in April 2022. Economists debate why, from measurement of capacity in high-tech industries to structural changes in manufacturing, but the practical point is that "normal" today is lower than it was decades ago.
Industrial output has barely grown since 2007. The output index was 102.4 in December 2007 and 104.0 in September 2026, an increase of under 2% in nearly 19 years, and its record high, 104.1, dates from September 2018. Over the same period real GDP grew about 45%. US output has shifted toward services, which IP does not measure.
How to read IP and utilization
- Watch manufacturing. Utilities follow the weather and mining follows energy prices and drilling. Manufacturing output is the cleanest read on the industrial cycle.
- Use three-month changes. Monthly moves of a few tenths are often reversed. A string of declines in manufacturing output says more than one bad month.
- Read utilization against its own history. Rates well above the long-run average have tended to coincide with bottlenecks and rising goods prices, and to encourage investment in new capacity; rates far below it point to slack, idle plants and little pressure to invest.
- Compare with surveys. The PMI's production index is meant to track IP; when the two diverge for months, one of them is misleading. See PMI explained: how purchasing managers’ indexes work.
- Remember the revisions. The Fed revises the latest months each release and the whole series in its annual revision.
Markets react to IP less than to payrolls or CPI, since much of the information is already in the earlier jobs report (factory hours) and the PMIs. A large surprise in manufacturing can still move industrial and materials stocks, and utilization feeds into views on inflation and on corporate investment. Industrials' typical place in the cycle is covered in How sectors behave across the business cycle: cyclicals vs. defensives.
Pitfalls and limitations
- A narrow slice. IP covers manufacturing, mining and utilities, a minority of the US economy. It can fall while GDP rises.
- Weather and energy. Hot summers and cold winters swing utilities output, and oil and gas drilling swings mining, by several percent in a month.
- Capacity is estimated. The utilization rate depends on the Fed's model of capacity, which is revised and may not capture new technologies well.
- A drifting average. Comparing today's utilization with 1960s or 1970s levels, or with the long-run average, overstates slack if "normal" has fallen.
- Revisions. The annual revision can change several years of history.
- Real, not dollars. IP measures volumes. A rise in factory revenue from higher prices does not show up as higher production.
IP and capacity utilization compared with neighbors
| Indicator | Publisher | Frequency | Type | Covers |
|---|---|---|---|---|
| Industrial production | Federal Reserve | Monthly | Hard, real | Output of factories, mines, utilities |
| Capacity utilization | Federal Reserve | Monthly | Hard, ratio | Output as a share of sustainable capacity |
| Manufacturing PMIs | ISM, S&P Global | Monthly | Survey (soft) | Breadth of change at manufacturers |
| Real GDP | BEA | Quarterly | Hard, real | All final output, goods and services |
| Manufacturing payrolls | BLS | Monthly | Hard | Jobs and hours in manufacturing |
The broad output measure is in GDP explained: how gross domestic product is measured and read, and how the NBER combines indicators like IP into recession dates is in What is a recession? How the NBER dates US recessions.
How Finabase does it
Finabase does not publish industrial production or capacity utilization data, and has no economic calendar. The chart and figures on this page are ours, built from the Federal Reserve's G.17 data distributed through the St. Louis Fed's FRED database and from the October 16, 2026 release.
The app shows the markets most tied to industry. The Industrials sector page lists the sector's S&P 500 stocks, such as machinery, aerospace and transport companies, with the sector index; the Sector page compares all 11 sectors, including materials, energy and utilities. Futures on crude oil, natural gas and copper, which move with industrial demand, are on the Futures page.
Frequently asked questions
What does industrial production measure?
It measures the real output of US manufacturing, mining and electric and gas utilities, as an index with 2017 = 100. It does not cover construction, agriculture or services.
What is a normal capacity utilization rate?
The Fed's long-run average for total industry over 1972–2025 is 79.4%. Readings well above it suggest tight capacity; readings well below suggest slack. September 2026's 76.9% was 2.5 points below the average.
When is the industrial production report released?
The Federal Reserve publishes the G.17 release around the middle of each month, covering the previous month. September 2026 data were released on October 16, 2026.
Does falling industrial production mean a recession?
Not necessarily. Industrial output fell 4.5% over the year to December 2015 without an economy-wide recession, because the slump was concentrated in mining and energy. Every recession since 1967 has, however, included a large fall in industrial production.
Why is industrial production barely higher than in 2007?
Because US growth since then has come mostly from services, which IP doesn't measure. The output index rose from 102.4 in December 2007 to 104.0 in September 2026, while real GDP grew about 45%.
Summary
The Federal Reserve's G.17 release measures the real output of factories, mines and utilities as an index (2017 = 100) and compares it with estimated sustainable capacity to give the utilization rate. In September 2026 output rose 0.7%, led by utilities, and utilization was 76.9%, below its 1972–2025 average of 79.4%. Industrial output is far more cyclical than GDP, with year-over-year falls of 15% in 2009 and 17% in 2020, and utilization hit a record low of 64.1% in April 2020. Read manufacturing first, look at several months, compare with the PMIs, and remember that industry is a shrinking slice of an economy that now runs mostly on services.
Sources and further reading
- Board of Governors of the Federal Reserve System — About G.17 Industrial Production and Capacity Utilization
- Board of Governors of the Federal Reserve System — G.17 Industrial Production and Capacity Utilization, September 2026 (October 16, 2026)
- Federal Reserve Board — Industrial production index, 2017 = 100 (FRED series INDPRO), downloaded 2026-10-30
- Federal Reserve Board — Capacity utilization, total industry (FRED series TCU), downloaded 2026-10-30
- BEA — Real GDP, chained 2017 dollars (FRED series GDPC1), downloaded 2026-10-30